Many significant and commercially important uses of modem computer technology relate to images and videos. These include image and video processing, image and video analysis and computer vision applications. In computer vision applications, such as, for example, object recognition and optical character recognition, it has been found that a separation of illumination and material aspects of an image can significantly improve the accuracy and speed of computer performance. Significant pioneer inventions related to the illumination and material aspects of an image are disclosed in U.S. Pat. No. 7,873,219 to Richard Mark Friedhoff, entitled Differentiation Of Illumination And Reflection Boundaries and U.S. Pat. No. 7,672,530 to Richard Mark Friedhoff et al., entitled Method And System For Identifying Illumination Flux In An Image (hereinafter the Friedhoff Patents).
In a first exemplary embodiment of the present invention, an automated, computerized method is provided for processing a video. According to a feature of the present invention, the method comprises the steps of providing a video file depicting a video, in a computer memory; scale separating the video file to generate a first scale separated video and a second scale separated video corresponding to the video; filtering at least one of the first scale-separated video and the second scale-separated video independent of the other of the first scale-separated video and the second scale-separated video to provide at least one independently filtered video; and compressing the at least one independently filtered video to provide at least one compressed video.
In a second exemplary embodiment of the present invention, a computer system is provided. The computer system includes a CPU and a memory storing a video file containing a video. The CPU is arranged and configured to execute a routine to scale separate the video file to generate a first scale separated video and a second scale separated video corresponding to the video, filter at least one of the first scale separated video and the second scale separated video independent of the other of the first scale separated video and the second scale separated video to provide at least one independently filtered video and compress the at least one independently filtered video to provide at least one compressed video.
In a third exemplary embodiment of the present invention, a computer program product, disposed on a non-transitory computer readable media is provided. The computer program product includes computer executable process steps operable to control a computer to: provide a video file depicting a video, in a computer memory, generate a first scale separated video and a second scale separated video corresponding to the video, filter at least one of the first scale separated video and the second scale separated video independent of the other of the first scale separated video and the second scale separated video to provide at least one independently filtered video and compress the at least one independently filtered video to provide at least one compressed video.
In a fourth exemplary embodiment of the present invention, an automated, computerized method is provided for processing a video. According to a feature of the present invention, the method comprises providing a video file depicting a video, in a computer memory; scale separating the video file to generate a detail component corresponding to the video; filtering the detail component to provide a filtered detail component; and compressing the filtered detail component to provide a compressed filtered detail component.
In a fifth exemplary embodiment of the present invention, an automated, computerized method is provided for processing a video. According to a feature of the present invention, the method comprises providing a video file depicting a video, in a computer memory; scale separating the video file to generate a level component corresponding to the video; filtering the level component to provide a filtered level component; and compressing the filtered level component to provide a compressed filtered level component.
In accordance with yet further embodiments of the present invention, computer systems are provided, which include one or more computers configured (e.g., programmed) to perform the methods described above. In accordance with other embodiments of the present invention, non-transitory computer readable media are provided which have stored thereon computer executable process steps operable to control a computer(s) to implement the embodiments described above. The present invention contemplates a computer readable media as any product that embodies information usable in a computer to execute the methods of the present invention, including instructions implemented as a hardware circuit, for example, as in an integrated circuit chip. The automated, computerized methods can be performed by a digital computer, analog computer, optical sensor, state machine, sequencer, integrated chip or any device or apparatus that can be designed or programmed to carry out the steps of the methods of the present invention.
a is a flow chart for identifying Type C token regions in the image file of
b is an original image used as an example in the identification of Type C tokens.
c shows Type C token regions in the image of
d shows Type B tokens, generated from the Type C tokens of
Referring now to the drawings, and initially to
Moreover, the computer system 10 includes an object database 24 storing information on various objects that can appear in the video files 18 stored in the memory 16. The information includes information on the material make-up and material reflectance colors for each object stored in the database 24. The object database is coupled to the CPU 12, as shown in
Alternatively, the CPU 12 can be implemented as a microprocessor embedded in a device such as, for example, the digital camera 14 or a robot. The CPU 12 can also be equipped with a real time operating system for real time operations related to videos, in connection with, for example, a robotic operation or an interactive operation with a user.
As shown in
In an image operation, the CPU 12 operates to analyze the RGB values of the pixels of images of stored video file 18 to achieve various objectives, such as, for example, to identify regions of an image that correspond to a single material depicted in a scene recorded in the video file 18. A fundamental observation underlying a basic discovery of the present invention, is that an image comprises two components, material and illumination. All changes in an image are caused by one or the other of these components. A method for detecting of one of these components, for example, material, provides a mechanism for distinguishing material or object geometry, such as object edges, from illumination and shadow boundaries.
Such a mechanism enables techniques that can be used to generate intrinsic images. Each of the intrinsic images corresponds to an original image, i.e., video frame, for example, an image depicted in an input video file 18. The intrinsic images include, for example, an illumination image, to capture the intensity and color of light incident upon each point on the surfaces depicted in the image, and a material reflectance image, to capture reflectance properties of surfaces depicted in the image (the percentage of each wavelength of light a surface reflects). The separation of illumination from material in the intrinsic images provides the CPU 12 with images optimized for more effective and accurate and efficient further processing.
For example, according to a feature of the present invention, the intrinsic images are applied in a digital image signal compression algorithm, for improved results in data transmission and/or storage. Computer files that depict an image, particularly a color image, require a significant amount of information arranged as, for example, pixels represented by bytes. Thus, each video file requires a significant amount of storage space in a memory, and can consume a large amount of time in a data transmission of the image to a remote site or device. The amount of time that can be required to transmit a sequence of images, for example, as in a video stream, can render an operation, such as a streaming operation for realtime display of a video on a smartphone, Internet website or tablet, unfeasible.
Accordingly, mathematical techniques have been developed to compress the number of bytes representing the pixels of an image to a significantly smaller number of bytes. For example, standards for lossy video compression developed by organizations such as ISO MPEG, the Moving Picture Experts Group, enable compression of digital video files. A compressed video can be stored in a manner that requires much less storage capacity than the original video file, and transmitted to a remote site or device in a far more efficient and speedy transmission operation. The compressed video file is decompressed for further use, such as, for example, display on a screen. However, due to the rapidly increasing number of users of devices for reception and realtime display of digital videos, known compression techniques are being pressed to the limits of effective functionality.
According to a feature of the present invention, digital signal compression and decompression processing is improved by performing the compression and decompression processes on intrinsic images.
Pursuant to a feature of the present invention, processing is performed at a token level. A token is a connected region of an image wherein the pixels of the region are related to one another in a manner relevant to identification of image features and characteristics such as an identification of materials and illumination. The pixels of a token can be related in terms of either homogeneous factors, such as, for example, close correlation of color among the pixels, or inhomogeneous factors, such as, for example, differing color values related geometrically in a color space such as RGB space, commonly referred to as a texture. The present invention utilizes spatio-spectral information relevant to contiguous pixels of images depicted in a video file 18 to identify token regions. The spatio-spectral information includes spectral relationships among contiguous pixels, in terms of color bands, for example the RGB values of the pixels, and the spatial extent of the pixel spectral characteristics relevant to a single material.
According to one exemplary embodiment of the present invention, tokens are each classified as either a Type A token, a Type B token or a Type C token. A Type A token is a connected image region comprising contiguous pixels that represent the largest possible region of the image encompassing a single material in the scene (uniform reflectance). A Type B token is a connected image region comprising contiguous pixels that represent a region of the image encompassing a single material in the scene, though not necessarily the maximal region of uniform reflectance corresponding to that material. A Type B token can also be defined as a collection of one or more image regions or pixels, all of which have the same reflectance (material color) though not necessarily all pixels which correspond to that material color. A Type C token comprises a connected image region of similar image properties among the contiguous pixels of the token, where similarity is defined with respect to a noise model for the imaging system used to record the image.
Referring now to
A 1st order uniform, homogeneous Type C token comprises a single robust color measurement among contiguous pixels of the image. At the start of the identification routine, the CPU 12 sets up a region map in memory. In step 100, the CPU 12 clears the region map and assigns a region ID, which is initially set at 1. An iteration for the routine, corresponding to a pixel number, is set at i=0, and a number for an N×N pixel array, for use as a seed to determine the token, is set an initial value, N=Nstart. Nstart can be any integer >0, for example it can be set at set at 11 or 15 pixels.
At step 102, a seed test is begun. The CPU 12 selects a first pixel, i=1, pixel (1, 1) for example (see
If the pixel value is at imax, a value selected as a threshold for deciding to reduce the seed size for improved results, the seed size, N, is reduced (step 110), for example, from N=15 to N=12. In an exemplary embodiment of the present invention, imax can be set at a number of pixels in an image ending at pixel (n, m), as shown in
After reduction of the seed size, the routine returns to step 102, and continues to test for token seeds. An Nstop value (for example, N=2) is also checked in step 110 to determine if the analysis is complete. If the value of N is at Nstop, the CPU 12 has completed a survey of the image pixel arrays and exits the routine.
If the value of i is less than imax, and N is greater than Nstop, the routine returns to step 102, and continues to test for token seeds.
When a good seed (an N×N array with approximately equal pixel values) is found (block 104), the token is grown from the seed. In step 112, the CPU 12 pushes the pixels from the seed onto a queue. All of the pixels in the queue are marked with the current region ID in the region map. The CPU 12 then inquires as to whether the queue is empty (decision block 114). If the queue is not empty, the routine proceeds to step 116.
In step 116, the CPU 12 pops the front pixel off the queue and proceeds to step 118. In step 118, the CPU 12 marks “good” neighbors around the subject pixel, that is neighbors approximately equal in color value to the subject pixel, with the current region ID. All of the marked good neighbors are placed in the region map and also pushed onto the queue. The CPU 12 then returns to the decision block 114. The routine of steps 114, 116, 118 is repeated until the queue is empty. At that time, all of the pixels forming a token in the current region will have been identified and marked in the region map as a Type C token.
When the queue is empty, the CPU 12 proceeds to step 120. At step 120, the CPU 12 increments the region ID for use with identification of a next token. The CPU 12 then returns to step 106 to repeat the routine in respect of the new current token region.
Upon arrival at N=Nstop, step 110 of the flow chart of
While each Type C token comprises a region of the image having a single robust color measurement among contiguous pixels of the image, the token may grow across material boundaries. Typically, different materials connect together in one Type C token via a neck region often located on shadow boundaries or in areas with varying illumination crossing different materials with similar hue but different intensities. A neck pixel can be identified by examining characteristics of adjacent pixels. When a pixel has two contiguous pixels on opposite sides that are not within the corresponding token, and two contiguous pixels on opposite sides that are within the corresponding token, the pixel is defined as a neck pixel.
If no, the CPU 12 exits the routine of
If yes, the CPU 12 proceeds to step 128 and operates to regrow the token from a seed location selected from among the unmarked pixels of the current token, as per the routine of
Subsequent to the regrowth of the token without the previously marked pixels, the CPU 12 returns to step 122 to test the newly regrown token. Neck testing identifies Type C tokens that cross material boundaries, and regrows the identified tokens to provide single material Type C tokens suitable for use in creating Type B tokens.
d shows Type B tokens generated from the Type C tokens of
A method and system for separating illumination and reflectance using a log chromaticity representation is disclosed in U.S. Pat. No. 7,596,266, which is hereby expressly incorporated by reference. The techniques taught in U.S. Pat. No. 7,596,266 can be used to provide illumination invariant log chromaticity representation values for each color of an image, for example, as represented by Type C tokens. Logarithmic values of the color band values of the image pixels are plotted on a log-color space graph. The logarithmic values are then projected to a log-chromaticity projection plane oriented as a function of a bi-illuminant dichromatic reflection model (BIDR model), to provide a log chromaticity value for each pixel, as taught in U.S. Pat. No. 7,596,266. The BIDR Model predicts that differing color measurement values fall within a cylinder in RGB space, from a dark end (in shadow) to a bright end (lit end), along a positive slope, when the color change is due to an illumination change forming a shadow over a single material of a scene depicted in the image.
Thus, according to the technique disclosed in U.S. Pat. No. 7,596,266, the RGB values of each pixel in each image of video file 18 can be mapped by the CPU 12 from the image file value p(n, m, R, G, B) to a log value, then, through a projection to the chromaticity plane, to the corresponding u, v value, as shown in
According to a feature of the present invention, log chromaticity values are calculated for each color depicted in an image of video file 18 input to the CPU 12 for identification of regions of the uniform reflectance (Type B tokens). For example, each pixel of a Type C token will be of approximately the same color value, for example, in terms of RGB values, as all the other constituent pixels of the same Type C token, within the noise level of the equipment used to record the image. Thus, an average of the color values for the constituent pixels of each particular Type C token can be used to represent the color value for the respective Type C token in the log chromaticity analysis.
Blend pixels are pixels between two differently colored regions of an image. If the colors between the two regions are plotted in RGB space, there is a linear transition between the colors, with each blend pixel, moving from one region to the next, being a weighted average of the colors of the two regions. Thus, each blend pixel does not represent a true color of the image. If blend pixels are present, relatively small Type C tokens, consisting of blend pixels, can be identified for areas of an image between two differently colored regions. By requiring a size minimum, the CPU 12 can eliminate tokens consisting of blend pixel from the analysis.
In step 204, the CPU 12 can alternatively collect colors at the pixel level, that is, the RGB values of the pixels of the input image of video file 18, as shown in
In this approach, the CPU 12 calculates a second derivative at each pixel, or a subset of pixels disbursed across the image to cover all illumination conditions of the image depicted in an input video file 18, using a Difference of Gaussians, Laplacian of Gaussian, or similar filter. The second derivative energy for each pixel examined can then be calculated by the CPU 12 as the average of the absolute value of the second derivative in each color band (or the absolute value of the single value in a grayscale image), the sum of squares of the values of the second derivatives in each color band (or the square of the single value in a grayscale image), the maximum squared second derivative value across the color bands (or the square of the single value in a grayscale image), or any similar method. Upon the calculation of the second derivative energy for each of the pixels, the CPU 12 analyzes the energy values of the pixels. There is an inverse relationship between second derivative energy and pixel stability, the higher the energy, the less stable the corresponding pixel.
In step 206, the CPU 12 outputs a list or lists of color (after executing one or both of steps 202 and/or 204). According to a feature of the present invention, all of the further processing can be executed using the list from either step 202 or 204, or vary the list used (one or the other of the lists from steps 202 or 204) at each subsequent step.
As taught in U.S. Pat. No. 7,596,266, and as noted above, alignment of the chromaticity plane is represented by N, N being a vector normal to the chromaticity representation, for example, the chromaticity plane of
For a higher dimensional set of colors, for example, an RYGB space (red, yellow, green, blue), the log chromaticity normal, N, defines a sub-space with one less dimension than the input space. Thus, in the four dimensional RYGB space, the normal N defines a three dimensional log chromaticity space. When the four dimensional RYGB values are projected into the three dimensional log chromaticity space, the projected values within the log chromaticity space are unaffected by illumination variation.
In step 214, the CPU 12 outputs an orientation for the normal N. As illustrated in the example of
In step 224, the CPU 12 operates to calculate a log value for each color in the list of colors and plots the log values in a three dimensional log space at respective (log R, log G, log B) coordinates, as illustrated in
According to a feature of the present invention, the CPU 12 then projects the log values for the colors A, B and C onto the log chromaticity plane to determine a u, v log chromaticity coordinate for each color. Each u, v log chromaticity coordinate can be expressed by the corresponding (log R, log G, log B) coordinate in the three dimensional log space. The CPU 12 outputs a list of the log chromaticity coordinates in step 226. The list cross-references each color to a u, v log chromaticity coordinate and to the pixels (or a Type C tokens) having the respective color (depending upon the list of colors used in the analysis (either step 202 (tokens) or 204 (pixels))).
In step 234, the CPU 12 optionally operates to augment each log chromaticity coordinate with a tone mapping intensity for each corresponding pixel (or Type C token). The tone mapping intensity is determined using any known tone mapping technique. An augmentation with tone mapping intensity information provides a basis for clustering pixels or tokens that are grouped according to both similar log chromaticity coordinates and similar tone mapping intensities. This improves the accuracy of a clustering step.
In step 236, the CPU 12 optionally operates to augment each log chromaticity coordinate with x, y coordinates for the corresponding pixel (or an average of the x, y coordinates for the constituent pixels of a Type C token) (see
In each of steps 234 and 236, the augmented information can, in each case, be weighted by a factor w1 and w2, w3 respectively, to specify the relative importance and scale of the different dimensions in the augmented coordinates. The weight factors w1 and w2, w3 are user-specified. Accordingly, the (log R, log G, log B) coordinates for a pixel or Type C token is augmented to (log R, log G, log B, T*w1, x*w2, y*w3) where T, x and y are the tone mapped intensity, the x coordinate and the y coordinate, respectively.
In step 238, the CPU 12 outputs a list of the augmented coordinates. The augmented log chromaticity coordinates provide accurate illumination invariant representations of the pixels, or for a specified regional arrangement of an input image, such as, for example, Type C tokens. According to a feature of the present invention, the illumination invariant characteristic of the log chromaticity coordinates is relied upon as a basis to identify regions of an image of a single material or reflectance, such as, for example, Type B tokens.
In step 244, the CPU 12 outputs a list of the cluster group memberships for the log chromaticity coordinates (cross referenced to either the corresponding pixels or Type C tokens) and/or a list of cluster group centers.
As noted above, in the execution of the clustering method, the CPU 12 can use the list of colors from either the list generated through execution of step 202 of the routine of
In step 252, the list of cluster centers is input to the CPU 12. In step 254, the CPU 12 operates to classify each of the log chromaticity coordinates identified in step 250, according to the nearest cluster group center. In step 256, the CPU 12 outputs a list of the cluster group memberships for the log chromaticity coordinates based upon the new list of colors, with a cross reference to either corresponding pixels or Type C tokens, depending upon the list of colors used in step 250 (the list of colors generated in step 202 or the list of colors generated in step 204).
In step 266, the CPU 12 operates to merge each of the pixels, or specified regions of an input image, such as, for example, Type C tokens, having a same cluster group membership into a single region of the image to represent a region of uniform reflectance (Type B token). The CPU 12 performs such a merge operation for all of the pixels or tokens, as the case may be, for the corresponding image of input video file 18. In step 268, the CPU 12 outputs a list of all regions of uniform reflectance (and also of similar tone mapping intensities and x, y coordinates, if the log chromaticity coordinates were augmented in steps 234 and/or 236). It should be noted that each region of uniform reflectance (Type B token) determined according to the features of the present invention, potentially has significant illumination variation across the region.
U.S. Patent Publication No. US 2010/0142825 teaches a constraint/solver model for segregating illumination and material in an image, including an optimized solution based upon a same material constraint. A same material constraint, as taught in U.S. Patent Publication No. US 2010/0142825, utilizes Type C tokens and Type B tokens, as can be determined according to the teachings of the present invention. The constraining relationship is that all Type C tokens that are part of the same Type B token are constrained to be of the same material. This constraint enforces the definition of a Type B token, that is, a connected image region comprising contiguous pixels that represent a region of the image encompassing a single material (same reflectance) in the scene, though not necessarily the maximal region corresponding to that material. Thus, all Type C tokens that lie within the same Type B token are by the definition imposed upon Type B tokens, of the same material, though not necessarily of the same illumination. The Type C tokens are therefore constrained to correspond to observed differences in appearance that are caused by varying illumination.
Since: ma=ia−la, mb=ib−lb, and mc=ic−lc these mathematical relationships can be expressed, in a same material constraint, as (1)la+(−1)lb+(0)lc=(ia−b), (1)la+(0)lb+(−1)lc=(ia−ic) and (0)la+(1)lb+(−1)lc=(ib−ic).
Thus, in the matrix equation of
Once the illumination values are known, the material color can be calculated by the CPU 12 using the I=ML equation. Intrinsic illumination and material images can be now be generated for the region defined by tokens a, b and c, by replacing each pixel in the original image by the calculated illumination values and material values, respectively. An example of an illumination image and material image, corresponding to the original image shown in
According to a feature of a further exemplary embodiment of the present invention, the CPU 12 is coupled to an object database 24. As noted above, the object database 24 stores a list of objects that can appear in the video files 18, and information on the material make-up and material reflectance colors for each object stored in the database 24. In connection with the above-described techniques for segregating an image into corresponding material reflectance and illumination intrinsic images, the CPU 12 is operated to perform a known object recognition task, such as, for example, a SIFT technique, to identify objects in an image being processed.
Upon the identification of an object in a scene depicted in an image being processed, the CPU 12 accesses the object database 24 for the material reflectance color information relevant to the identified object. The CPU 12 is then operated to correlate, for example, any Type C tokens in the image being processed that constitute the identified object. The material reflectance color information for the identified object can then be used to specify, for example, a fixed material color anchor value added to the matrix equation shown in
According to yet another feature of the exemplary embodiment, the CPU 12 is coupled to the Internet 26. In this manner, the CPU 12 can access websites 28 on the Internet 26. The websites 28 provide another source for an object database. For example, the CPU 12 can search the Internet 26 via, for example, a text-based search, to obtain information at an accessed website 28, relevant to the material characteristics of an object identified in an image being processed. The material characteristics are used to determine the fixed anchor value described above.
Implementation of the constraint/solver model according to the techniques and teachings of U.S. Patent Publication No. US 2010/0142825, utilizing, for example, the Type C tokens and Type B tokens obtained, for example, via a log chromaticity clustering technique according to the present invention, and information from an object database 26, provides a highly effective and efficient method for generating intrinsic images corresponding to an original input image. The intrinsic images can be used to enhance the accuracy, speed and efficiency of image processing, image analysis and computer vision applications.
According to yet another feature of the present invention, advantage is made of a correspondence between inherent characteristics of each of the intrinsic material reflectance and illumination images with observations of human visual perception. As observed, human perception of details of objects depicted in a scene recorded in an video file 18 is aligned with the details depicted in the intrinsic images for the material reflectance aspects of the scene. Moreover, human perception of motion depicted in a sequence of images for the scene is aligned with motion displayed in a sequence of intrinsic images for the illumination aspects of the scene.
Humans tend to perceive fine spatial detail with more clarity in static or slow-moving regions of a video and tend to perceive fast motion more clearly in larger spatial objects or regions of a video. In order to allow perception of both the fine details and the fast motion, conventional video compression techniques maintain high frame rates to allow for perception of smooth motion and high spatial resolution for perception of fine detail.
Embodiments of the present invention allow the material component and the illumination component of a video to be separated from each other in a precompression technique into an independent material video and an independent illumination video for filtering. Such separation of the material and illumination videos allows adjustments to be made to the material and illumination video frames making up the video independently of each other for further reduction in video file size, yet maintaining aspects of the original video frames that are most important for human perception of videos. Because videos are formed of sequential images, it is possible to alter or remove individual video frames of the video without affecting the quality of the video from a human perception standpoint. Due to the importance of material reflectance of an image for fine details and object boundaries in a video, but not necessarily the shape and movement, it is possible to reduce the frame rate of the material images for storage or transmission without affecting the quality of the video from a human perception standpoint. Also, due to the importance of illumination of an image for the shape and movement in a video, but not necessarily the fine details and object boundaries, it is possible to reduce the detail of the illumination images storage or transmission without affecting the quality of the video from a human perception standpoint.
In step 500, the CPU 12 receives an original video file, for example, a video file 18 from the memory 16. In step 502, the CPU 12 operates to generate intrinsic images from the each of the video frames of the original video file, for example, according to the techniques described in detail above, to output illumination maps (illumination video frames forming an illumination video) (step 504) and reflectance maps (material video frames forming a material video) (step 506).
In step 508, the CPU 12 operates to separately perform, either in a parallel operation, or in a sequence, an illumination component filtering on the illumination video frames in step 510 and a material component filtering on the material video frames in step 512. In this embodiment, the illumination component filtering in step 510 includes spatially subsampling the illumination video and the material component filtering in step 512 includes temporally subsampling the material video. The spatial subsampling of the illumination video may include reducing the spatial resolution of each of the illumination video frames of the illumination video. For example, the spatial resolution of illumination video frames may reduced both horizontally and vertically by a factor of two, such that a spatial resolution W×H of the illumination video frames is reduced to W/2×H/2 while not affecting the frame rate F. The spatial resolution of the illumination video frames of the illumination video may also be decreased during the spatial subsampling by other amounts in other examples. The temporal subsampling of the material video may include removing j material video frame(s) out of every k material video frames of the material video in a repeating pattern. For example, where j=1 and k=2, every other material video frame is removed from the material video, in a repeating pattern of removing the first video frame of each group of two video frames and leaving the second video frame of the group of two video frames or in a repeating pattern of removing the second video frame of each group of two video frames and leaving the first video frame of the group of two video frames.
Also, for example, where j=2 and k=3, two out of every three material video frames may be removed from the material video during the temporal subsampling in a first repeating pattern where the first and second material video frames of each group of three material video frames are removed and the third material video frame of the group of three material video frames is not removed, a second repeating pattern where the first and third material video frames of each group of three material video frames are removed and the second material video frame of the group of three material video frames is not removed, or a third repeating pattern where the second and third material video frames of each group of three material video frames are removed and the first material video frame of the group of three material video frames is not removed.
The foregoing examples are merely illustrative and the number of material video frames of the material video removed and/or the pattern of removal may also be varied during the temporal subsampling by other amounts in other examples.
The spatial subsampling and the temporal subsampling reduce the sizes of the illumination video and the material video, reducing the size of the video file storing the illumination and material videos. In step 510, the CPU 12 may perform one or more alternative or additional filtering processes on each of the illumination video frames and in step 512, the CPU 12 may perform one or more alternative or additional filtering processes on each of the material video frames.
In a step 514, the CPU 12 operates to separately interpolate the filtered illumination video and the filtered material video and then re-mix the interpolated illumination video and the interpolated material video according to a pixel-by-pixel or sample-by-sample operation to form a recombined intrinsic video. CPU 12 or the remote device operates to perform, either in a parallel operation, or in a sequence, separate interpolation processes on the filtered illumination video and the filtered reflectance video. In this embodiment, the file size of the interpolated illumination video and the interpolated material video are reduced compared the corresponding illumination video and material video created in step 508.
The interpolating may include creating interpolated illumination frames from the filtered illumination frames created in the illumination component subsampling in step 508. The interpolated illumination frames may be formed by interpolating spatially between pairs of horizontally and vertically adjacent pixels of each of the filtered illumination frames created in step 510 to output an interpolated illumination video (step 532). For example, referring to
The interpolating may also include creating interpolated material frames to replace the material frames removed in the material component subsampling in step 512. The interpolated material frames may be formed by interpolating each pixel position of a material frame directly preceding the corresponding removed material frame and a material frame directly following the corresponding removed material frame to output an interpolated material video. For example, referring to
In alternative embodiments, other known methods of interpolation, for example linear interpolation, bilinear interpolation, cubic interpolation or bicubic interpolation can be used in step 514.
In a step 516, gamma correction and/or tone adjustment may be performed on the recombined intrinsic video. In a step 518, the recombined intrinsic video is compressed or encoded for transmission or storage. An encoder (or CPU carrying out the process) proceeds to compress or encode the recombined intrinsic video according to a known compression format such as H.264/AVC, HEVC or another format.
According to a feature of the present invention, in step 520, the compressed recombined intrinsic video (video formed of filtered and interpolated intrinsic images) is stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
In a step 524, a decoder of the CPU 12 or the remote device operates to decompress or decode the compressed recombined intrinsic video.
Each of steps 522 and 524 are implemented using known techniques for compression or decompression of digital video material, such as techniques compatible with one of ISO/MPEG-2 Visual, ITU-T H.264/AVC, HEVC or other known formats for compressed video material.
In step 526, the CPU 12 or the remote device operates to output a video appearing to the human visual system to be of essentially the same video quality as the original video, for example, the video depicted in the video file 18 initially processed by the CPU 12 according to the routine of
Steps 600, 602, 604, 606 of
Steps 608, 610, 612 of
Starting at step 614, the method of
In a step 620, the CPU 12 operates to separately compress or encode, either in a parallel operation, or in a sequence, filtered illumination video and the filtered material video, which are performed by separate encoders 620a, 620b, respectively, of CPU 12. For example, the CPU 12 operates to convert the illumination maps to a known sampling format such as RGB, YCrCb or YUV. The CPU 12 then proceeds to compress the converted illumination maps and reflectance maps according to a known compression format such as H.264/AVC, HEVC or another format. The individual encoders 620a, 620b may optionally communicate with each other while compressing the filtered illumination video and the filtered material video, respectively. In one embodiment, steps 610, 612 and/or steps 616, 618 may also be performed by encoders 620a, 620b.
According to a feature of the present invention, in step 622, the compressed filtered illumination video (video formed of filtered and compressed illumination images) and the compressed filtered material video (video formed of filtered and compressed material images), either in a parallel operation, or in a sequence, are stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
In steps 626 and 628, in contrast to step 524 of
In decompression process of step 626, decoder 620a performs a decompression process on the compressed version of the illumination video to output the decompressed filtered illumination video.
In decompression process of step 628, decoder 620b performs a decompression process on the compressed version of the material video to output the decompressed filtered reflectance video.
Each of steps 624, 626 and 628 are implemented using known techniques for compression or decompression of digital video material, such as techniques compatible with one of ISO/MPEG-2 Visual, ITU-T H.264/AVC, HEVC or other known formats for compressed video material.
In steps 630, 632, in a similar same manner as in step 514 of figure, CPU 12 or the remote device operates to perform, either in a parallel operation, or in a sequence, a spatial interpolation process on the filtered illumination video and temporal interpolation process on the filtered reflectance video.
Step 630 may include creating interpolated illumination frames from the filtered illumination frames created in the illumination component subsampling in step 610. The interpolated illumination frames by interpolating spatially between pairs of horizontally and vertically adjacent pixels of each of the filtered illumination frames created in step 610 to output an interpolated illumination video (step 634). Step 630 results in an illumination video including a sequence of illumination frames at the original resolution, frame rate F and spatial dimensions W×H.
Step 632 may include creating interpolated material frames to replace the material frames removed in the material component subsampling in step 612. The interpolated material frames may be formed by interpolating each pixel position of a material frame directly preceding the corresponding removed material frame and a material frame directly following the corresponding removed material frame to output an interpolated material video (step 636). Step 632 results in a material video including a sequence of material frames at the original resolution W×H and frame rate F.
In step 638, the CPU 12 or the remote device operates to recombine the illumination video output at step 634 and the material video output at step 636 to output a video appearing to the human visual system to be of essentially the same video quality as the original video (step 640), for example, the video depicted in the video file 18 initially processed by the CPU 12 according to the routine of
Steps 700, 702, 704, 706 of
In step 708, the CPU 12 operates to separately perform, either in a parallel operation, or in a sequence, an illumination component filtering on the illumination video frames in a step 710 and a material component filtering on the material video frames in a step 712. In this embodiment, in contrast with the methods of
The filtering reduce the sizes of the illumination video and the material video. Filters may be properly chosen such that the size reduction and quality performance is adjusted to be essentially identical to the method described with respect to
Steps 714, 716, 718, 720, 722 of
In a step 720, the CPU 12 operates to separately compress or encode, either in a parallel operation, or in a sequence, filtered illumination video and the filtered material video, which are performed by separate encoders 720a, 720b, respectively, or CPU 12. For example, the CPU 12 operates to convert the illumination maps to a known sampling format such as RGB, YCrCb or YUV. The CPU 12 then proceeds to compress the converted illumination maps and reflectance maps according to a known compression format such as H.264/AVC, HEVC or another format. The individual encoders 720a, 720b may optionally communicate with each other while compressing the filtered illumination video and the filtered material video, respectively.
According to a feature of the present invention, in step 722, the compressed filtered illumination video (video formed of filtered and compressed illumination images) and the compressed filtered material video (video formed of filtered and compressed material images), either in a parallel operation, or in a sequence, are stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
In step 724, depending on whether the compressed recombined filtered intrinsic video is stored or transmitted in step 722, the compressed recombined filtered intrinsic video is retrieved by CPU 12 or received by the remote device as a website 28 via the Internet 26.
In steps 726 and 728, decoders 720a, 720b of the CPU 12 or the remote device operate to perform, either in a parallel operation, or in a sequence, decompression (decoding) processes.
In decompression process of step 726, decoder 720a performs a decompression process on the compressed version of the illumination video to output the decompressed filtered illumination video (step 730).
In decompression process of step 728, decoder 720b performs a decompression process on the compressed version of the material video to output the decompressed filtered reflectance video (732).
Each of steps 724, 726 and 728 are implemented using known techniques for compression or decompression of digital video material, such as techniques compatible with one of ISO/MPEG-2 Visual, ITU-T H.264/AVC, HEVC or other known formats for compressed video material.
In step 734, the CPU 12 or the remote device operates to recombine the illumination video output at step 730 and the material video output at step 732 to output a video appearing to the human visual system to be of essentially the same video quality as the original video (step 736), for example, the video depicted in the video file 18 initially processed by the CPU 12 according to the routine of
Steps 800, 802, 804, 806 of
Steps 808, 810, 812 are the same as the steps 708, 710, 712 of
Spatial and temporal filters may be properly chosen such that the reduction in size and quality performance is adjusted to be essentially identical to the method described with respect to
In a step 814, the CPU 12 operates to re-mix the filtered illumination video and the filtered material video according to a pixel-by-pixel or sample-by-sample operation to form a recombined filtered intrinsic video including both the filtered illumination video frames and the filtered material video frames.
In a step 816, the CPU 12 may operate to separately perform gamma correction and/or tone adjustment on the recombined intrinsic video.
In a step 818, an encoder of CPU 12 compresses or encodes the recombined filtered intrinsic video for transmission or storage. The encoder proceeds to compress or encode the recombined filtered intrinsic video according to a known compression format such as H.264/AVC, HEVC or another format.
According to a feature of the present invention, in step 820, the compressed recombined filtered intrinsic video (video formed of filtered intrinsic images) is stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
In a step 824, a decoder the CPU 12 or the remote device operates to perform a decompression or decoding process on the compressed recombined filtered intrinsic video to output the recombined video (step 826). The decompression or decoding is implemented using known techniques for compression or decompression of digital video material, such as techniques compatible with one of ISO/MPEG-2 Visual, ITU-T H.264/AVC, HEVC or other known formats for compressed video material. The output video appears to the human visual system to be of essentially the same video quality as the original video (step 736), for example, the video depicted in the video file 18 initially processed by the CPU 12 according to the routine of
According to yet another feature of the present invention, instead of segregating a video file into illumination and material videos, the video file is subject to a scale separation operation. Scale separation is a technique for separating local variation within an image from global variation. An image is separated into large-scale features and small-scale features. A known method for performing scale separation on an image is described in “Fast Bilateral Filtering for the Display of High-Dynamic-Range Images,” Frédo Durand and Julie Dorsey, ACM Transactions of Graphics (Proceedings of the ACM SIGGRAPH '02 Conference). Durand and Dorsey describe as scale separation technique that uses a bilateral filter to separate an image into a “level” channel and a “detail” channel. The level channel includes the low frequency components of the image and depicts large scale variations of the image, without details, which are depicted in the detail channel as high frequency components of the image. As such, the level channel is a reasonable approximation of log illumination intensity of the image, and the detail channel is a reasonable approximation of the log material intensity.
It is also known that bilateral filtering can be applied to videos, as is discussed in “Seperable bilateral filtering for fast video preprocessing,” Tuan Q. Pham and Lucas J. van Vliet, International Conference on Multimedia and Expo, IEEE, 2005. Pham uses a bilateral filter to reduce the noise in a video before compression, which can improve video compression. The bilateral filter operates temporally as well as in the spatial and intensity dimensions. By blurring away fine details (much of which is noise) but leaving the important large structures of the video, the resulting video stream is a smaller file at the same quality or a file of the same size with higher quality.
Another method of scale separation is described in “Two methods for display of high contrast images,” Jack Tumblin and Jessica K. Hodgins, ACM Trans. on Graphics 18, 1, pages 56-94, which describes using a Guassian blur filtering operation that can be used to separate out large-scale features and small-scale features.
With this feature of the present invention, advantage is made of a correspondence between low frequency and high frequency scale separation channels with observations of human visual perception. As observed, human perception of details of objects depicted in a scene recorded in a video file 18 is aligned with the details depicted in high frequency data provided by scale separation operations. Moreover, human perception of motion depicted in a sequence of images for the scene is aligned with motion displayed in a sequence of images formed from low frequency data provided by scale separation operations.
Embodiments of the present invention allow the high frequency component and the low frequency component of a video to be separated from each other in a precompression technique into separate components for filtering. Such separation of the high and low frequency components allows adjustments to be made to the high and low frequency components making up the video independently of each other for further reduction in video file size, yet maintaining aspects of the original video frames that are most important for human perception of videos. It is possible to alter or remove individual video frames of the video without affecting the quality of the video from a human perception standpoint. Due to the importance of high frequency components of an image for fine details and object boundaries in a video, but not necessarily the shape and movement, it may be possible to reduce the frame rate of the high frequency components for storage or transmission without affecting the quality of the video from a human perception standpoint. Also, due to the importance of low frequency components of an image for the shape and movement in a video, but not necessarily the fine details and object boundaries, it is possible to reduce the detail of the low frequency components for storage or transmission without affecting the quality of the video from a human perception standpoint.
The video processing method shown in
In step 900, the CPU 12 receives an original video file, for example, a linear video file 18 from the memory 16. In step 902, the CPU 12 operates to scale separate the video file to output low frequency components—the larger structures—in a level video (step 904) and high frequency components—the details—in a detail video (step 906). In one preferred embodiment, which is based on the method disclosed in “Fast Bilateral Filtering for the Display of High-Dynamic-Range Images,” Frédo Durand and Julie Dorsey, ACM Transactions of Graphics (Proceedings of the ACM SIGGRAPH '02 Conference), the scale separation in step 902 includes applying a bilateral filter implementation to the video file to generate the level channel (the low frequency components representing the large scale features), then the level channel is subtracted from the original image to generate the detail channel (the high frequency components representing the fine details). The bilateral filter can be the temporal bilateral filter disclosed in “Seperable bilateral filtering for fast video preprocessing,” Tuan Q. Pham and Lucas J. van Vliet, International Conference on Multimedia and Expo, IEEE, 2005, but can be applied to the video in the manner described by Durand and Dorsey to separate the video into the large scale features (approximation of illumination) and the fine details (approximation of reflectance). Accordingly, the temporal bilateral filter can be applied to generate the level video, then the level video is subtracted from the original video to generate the detail video. The temporal bilateral filter can be applied with a larger range sigma and spatial sigma (as described by Durand and Dorsey) than it would be in a noise reduction technique.
The level video and the detail video can be calculated as the bilateral blur of the intensity image or as the bilateral blur of each of the R, G and B independently. If the level video and detail video are calculated from the intensity image, the color from the original image is recombined with the level video and the detail video after they are filtered and recombined. If the level video and detail video are calculated from each of the R, G and B independently, the level video and detail video include the color components and it is not necessary to recombine the color from the original image with the level video and the detail video after they are filtered and recombined.
Because the video file is a linear video file, step 902 includes putting the video file through a log transform before the temporal bilateral filter is applied. In an alternative embodiment, the log transform may be replaced by a gamma correction operation, which behaves very similarly to a log transform operation.
The exemplary embodiment of scale separation involves bilateral filtering in the log domain; however, in other embodiments of the present invention, the scale separation can be performed by using any blurring filter, for example a Gaussian filter. The blurring can be performed in any domain, for example linear, log or gamma corrected. Performance will be better with any filter of the class of “edge preserving blurring filters,” such as bilateral filters, median filters, anisotropic diffusion, or guided filters, as described in “Guided Image Filtering,” K. He, J. Sun and X. Tang, Proceeding of European Conference Computer Vision (ECCV) (2010).
In step 908, the CPU 12 operates to separately perform, either in a parallel operation, or in a sequence, a level component filtering on the level video output in step 910 and a detail component filtering on the detail video output in step 912. In this embodiment, the level component filtering in step 910 includes spatially subsampling the level video and the detail component filtering in step 912 includes temporally subsampling the detail video. In this embodiment, similar to the subsampling described above with respect to
The spatial subsampling and the temporal subsampling reduce the sizes of the level video and the detail video, reducing the size of the video file storing the level and detail videos. In step 910, the CPU 12 can perform one or more alternative or additional filtering processes on each of the level video frames, and in step 912, the CPU 12 can perform one or more alternative or additional filtering processes on each of the detail video frames.
In a step 914, the CPU 12 operates to separately interpolate the filtered level video and the filtered detail video and then re-mix the interpolated level video and the interpolated detail video according to a pixel-by-pixel or sample-by-sample operation to form a recombined scale-separated video. In this embodiment, CPU 12 operates to perform, either in a parallel operation, or in a sequence, separate interpolation processes on the filtered level video and the filtered detail video. If the level video and detail video are in the log domain, the re-mixing involves element-wise adding all of the channels of the pixels of the level video and the detail video (log(video)=log(level)+log(detail)), then exponentiating the log space output to get back to the linear-space version of the video. In this embodiment, the recombined scale-separated video file has a reduced file size compared to the video file input at step 900.
Similar to step 514 of
In a step 916, gamma correction and/or tone adjustment can be performed on the recombined scale-separated video. In a step 918, the recombined scale-separated video is compressed or encoded for transmission or storage. An encoder (or CPU carrying out the process) proceeds to compress or encode the recombined scale-separated video according to a known compression format such as H.264/AVC, HEVC or another format.
According to a feature of the present invention, in step 920, the compressed recombined scale-separated video (video formed of filtered, interpolated and scale-separated images) is stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
In contrast to the method of
In step 1010, as with step 908, the CPU 12 operates to separately perform, either in a parallel operation, or in a sequence, a level component filtering on the level video output in step 1012, which includes spatially subsampling the level video, and a detail component filtering on the detail video output in step 1014, which includes temporally subsampling the detail video. Steps 1012 and 1014 can also include additional or alternative filtering operations.
In a step 1016, as with step 914, the CPU 12 operates to separately interpolate the filtered level video and the filtered detail video and then re-mix the interpolated level video and the interpolated detail video according to a pixel-by-pixel or sample-by-sample operation to form a recombined scale-separated video. In this embodiment, CPU 12 operates to perform, either in a parallel operation, or in a sequence, separate interpolation processes on the filtered level video and the filtered detail video. The re-mixing involves element-wise adding all of the channels of the pixels of the level video and the detail video (recombined video=interpolated level video+interpolated detail video).
Similar to step 914 of
In a step 1018, as with step 918, the recombined scale-separated video is compressed or encoded for transmission or storage. Because gamma correction/tone adjustment was applied to the video file in step 1002, gamma correction/tone adjustment does not need to be applied to the video before compression as with step 916 in
According to a feature of the present invention, in step 1020, as with step 920, the compressed recombined scale-separated video (video formed of filtered, scale-separated and interpolated images) is stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
Steps 1108, 1110, 1112 of
Starting at step 1114, the method of
In a step 1120, the CPU 12 operates to separately compress or encode, either in a parallel operation, or in a sequence, filtered level video and the filtered detail video, which are performed by separate encoders 1120a, 1120b, respectively, of CPU 12. For example, the CPU 12 operates to convert the level maps and detail maps to a known sampling format such as RGB, YCrCb or YUV. The CPU 12 then proceeds to compress the converted level maps and detail maps according to a known compression format such as H.264/AVC, HEVC or another format. The individual encoders 1120a, 1120b can optionally communicate with each other while compressing the filtered level video and the filtered detail video, respectively. In one embodiment, steps 1110, 1112 and/or steps 1116, 1118 can also be performed by encoders 1120a, 1120b.
According to a feature of the present invention, in step 1122, the compressed filtered level video (video formed of filtered and compressed level images) and the compressed filtered detail video (video formed of filtered and compressed detail images), either in a parallel operation, or in a sequence, are stored by the CPU 12 in the memory 16 and/or transmitted, for example, via the Internet 26, to a remote device configured, for example, as a website 28 (see
Steps 1210, 1212, 1214 of
Starting at step 1216, the method of
Then in step 1218, as with step 1122 of
In alternative embodiments of the methods described with respect to
In the preceding specification, the invention has been described with reference to specific exemplary embodiments and examples thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the claims that follow. The specification and drawings are accordingly to be regarded in an illustrative manner rather than a restrictive sense.