Apparatus and methods of obtaining multi-scale feature vector using CNN based integrated circuits

Abstract

A pixel feature vector extraction system for extracting multi-scale features contains a cellular neural networks (CNN) based integrated circuit (IC) for extracting pixel feature vector out of input imagery data by performing convolution operations using pre-trained filter coefficients of ordered convolutional layers in a deep learning model. The ordered convolutional layers are organized in a number of groups with each group followed by a pooling layer. Each group is configured for a different size of feature map. Pixel feature vector contains a combination of feature maps from at least two groups, for example, concatenation of the feature maps. The first group of the at least two groups contains the largest size of the feature maps amongst all of the at least two groups. Feature maps of the remaining of the at least two groups are modified to match the size of the feature map of the first group.

Description

FIELD

The patent document relates generally to the field of machine learning. More particularly, the present document relates to apparatus and methods of obtaining multi-scale feature vector (i.e., pixel feature vector) for image segmentation, object detection and optical flow.

BACKGROUND

Machine learning is an application of artificial intelligence. In machine learning, a computer or computing device is programmed to think like human beings so that the computer may be taught to learn on its own. The development of neural networks has been key to teaching computers to think and understand the world in the way human beings do. One of the neural networks is referred to as Cellular Neural Networks or Cellular Nonlinear Networks (CNN), which have been applied to many different fields and problems including, but not limited to, image processing since 1988. In recent years, CNN-based deep feature descriptors have been shown to be effective global descriptors for image retrieval. However, deep feature descriptors cannot be used for typical object detection, image segmentation and optical flow applications due to lack of local features in global descriptors.

SUMMARY

This section is for the purpose of summarizing some aspects of the invention and to briefly introduce some preferred embodiments. Simplifications or omissions in this section as well as in the abstract and the title herein may be made to avoid obscuring the purpose of the section. Such simplifications or omissions are not intended to limit the scope of the invention.

Pixel feature vector extraction systems for extracting multi-scale features are disclosed. In one aspect of the disclosure, a pixel feature vector extraction system contains a cellular neural networks (CNN) based IC configured for extracting a pixel feature vector out of an input imagery data by performing convolution operations using pre-trained filter coefficients of ordered convolutional layers in a deep learning model. The ordered convolutional layers are organized in a number of groups with each group followed by a pooling layer. Each group is configured for a different size of feature map. Pixel feature vector contains a combination of feature maps from at least two groups, for example, concatenation of the feature maps. The first group of the at least two groups contains the largest size of the feature maps amongst all of the at least two groups. Feature maps of the remaining of the at least two groups are modified to match the size of the feature map of the first group.

In another aspect of the disclosure, the CNN based IC includes a number of cellular neural networks (CNN) processing engines operatively coupled to at least one input/output data bus. The CNN processing engines are connected in a loop with a clock-skew circuit. Each CNN processing engine contains a CNN processing block configured for simultaneously obtaining results of the convolution operations, a first set of memory buffers for storing the input imagery data, and a second set of memory buffers for storing the pre-trained filter coefficients.

Objects, features, and advantages of the invention will become apparent upon examining the following detailed description of an embodiment thereof, taken in conjunction with the attached drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other features, aspects, and advantages of the invention will be better understood with regard to the following description, appended claims, and accompanying drawings as follows:

FIGS. 1A-1B are block diagrams illustrating an example integrated circuit designed for extracting features from input imagery data in accordance with one embodiment of the invention;

FIG. 2 is a function block diagram showing an example controller configured for controlling operations of one or more CNN processing engines according to an embodiment of the invention;

FIG. 3 is a diagram showing an example CNN processing engine in accordance with one embodiment of the invention;

FIG. 4 is a diagram showing M×M pixel locations within a (M+2)-pixel by (M+2)-pixel region, according to an embodiment of the invention;

FIGS. 5A-5C are diagrams showing three example pixel locations, according to an embodiment of the invention;

FIG. 6 is a diagram illustrating an example data arrangement for performing 3×3 convolutions at a pixel location, according to one embodiment of the invention;

FIG. 7 is a function block diagram illustrating an example circuitry for performing 3×3 convolutions at a pixel location, according to one embodiment of the invention;

FIG. 8 is a diagram showing an example rectification according to an embodiment of the invention;

FIGS. 9A-9B are diagrams showing two example 2×2 pooling operations according to an embodiment of the invention;

FIG. 10 is a diagram illustrating a 2×2 pooling operation reduces M-pixel by M-pixel block to a (M/2)-pixel by (M/2)-pixel block in accordance with one embodiment of the invention;

FIGS. 11A-11C are diagrams illustrating examples of M-pixel by M-pixel blocks and corresponding (M+2)-pixel by (M+2)-pixel region in an input image, according to one embodiment of the invention;

FIG. 12 is a diagram illustrating an example of a first set of memory buffers for storing received imagery data in accordance with an embodiment of the invention;

FIG. 13A is a diagram showing two operational modes of an example second set of memory buffers for storing filter coefficients in accordance with an embodiment of the invention;

FIG. 13B is a diagram showing example storage schemes of filter coefficients in the second set of memory buffers, according to an embodiment of the invention;

FIG. 14 is a diagram showing a plurality of CNN processing engines connected as a loop via an example clock-skew circuit in accordance of an embodiment of the invention;

FIG. 15 is a schematic diagram showing an example image processing technique based on convolutional neural networks in accordance with an embodiment of the invention;

FIG. 16 is a flowchart illustrating an example process of achieving a trained convolutional neural networks model having bi-valued 3×3 filter kernels in accordance with an embodiment of the invention;

FIG. 17 is a diagram showing an example filter kernel conversion scheme in accordance with the invention;

FIG. 18 is a diagram showing an example data conversion scheme in accordance with one embodiment of the invention;

FIG. 19 is a schematic diagram illustrating an example process of image processing for object detections, image segmentations or optical flow tasks, according to an embodiment of the invention;

FIG. 20 is a schematic diagram showing an example process of creating a pixel feature vector in accordance with one embodiment of the invention;

FIG. 21 is a schematic diagram showing an example pixel feature vector created from a VGG16 in accordance with an embodiment of the invention;

FIG. 22 is a schematic diagram showing an example task specific network for object detection using a pixel feature vector as an input in accordance with an embodiment of the invention;

FIG. 23 is a diagram showing an example image with an object detected in accordance with one embodiment of the invention; and

FIG. 24 is a diagram showing an example image processed with image segmentation in accordance with an embodiment of the invention.

DETAILED DESCRIPTIONS

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will become obvious to those skilled in the art that the invention may be practiced without these specific details. The descriptions and representations herein are the common means used by those experienced or skilled in the art to most effectively convey the substance of their work to others skilled in the art. In other instances, well-known methods, procedures, and components have not been described in detail to avoid unnecessarily obscuring aspects of the invention.

Reference herein to “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the invention. The appearances of the phrase “in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Further, the order of blocks in process flowcharts or diagrams or circuits representing one or more embodiments of the invention do not inherently indicate any particular order nor imply any limitations in the invention. Used herein, the terms “upper”, “lower”, “diagonal”, “off-diagonal”, “top”, “bottom”, “right” and “left” are intended to provide relative positions for the purposes of description, and are not intended to designate an absolute frame of reference

Embodiments of the invention are discussed herein with reference to FIGS. 1A-1B-FIGS. 20A-20B. However, those skilled in the art will readily appreciate that the detailed description given herein with respect to these figures is for explanatory purposes as the invention extends beyond these limited embodiments.

Referring first to FIG. 1A, it is shown a block diagram illustrating an example digital integrated circuit (IC) 100 for extracting features out of an input image in accordance with one embodiment of the invention.

The integrated circuit 100 is implemented as a digital semi-conductor chip and contains a CNN processing engine controller 110, and one or more neural networks (CNN) processing engines 102 operatively coupled to at least one input/output (I/O) data bus 120. Controller 110 is configured to control various operations of the CNN processing engines 102 for extracting features out of an input image based on an image processing technique by performing multiple layers of 3×3 convolutions with rectifications or other nonlinear operations (e.g., sigmoid function), and 2×2 pooling operations. To perform 3×3 convolutions requires imagery data in digital form and corresponding filter coefficients, which are supplied to the CNN processing engine 102 via input/output data bus 120. It is well known that digital semi-conductor chip contains logic gates, multiplexers, register files, memories, state machines, etc.

According to one embodiment, the digital integrated circuit 100 is extendable and scalable. For example, multiple copy of the digital integrated circuit 100 can be implemented on one semiconductor chip.

All of the CNN processing engines are identical. For illustration simplicity, only few (i.e., CNN processing engines 122a-122h, 132a-132h) are shown in FIG. 1B. The invention sets no limit to the number of CNN processing engines on a digital semi-conductor chip.

Each CNN processing engine 122a-122h, 132a-132h contains a CNN processing block 124, a first set of memory buffers 126 and a second set of memory buffers 128. The first set of memory buffers 126 is configured for receiving imagery data and for supplying the already received imagery data to the CNN processing block 124. The second set of memory buffers 128 is configured for storing filter coefficients and for supplying the already received filter coefficients to the CNN processing block 124. In general, the number of CNN processing engines on a chip is 2ⁿ, where n is an integer (i.e., 0, 1, 2, 3, . . . ). As shown in FIG. 1B, CNN processing engines 122a-122h are operatively coupled to a first input/output data bus 130a while CNN processing engines 132a-132h are operatively coupled to a second input/output data bus 130b. Each input/output data bus 130a-130b is configured for independently transmitting data (i.e., imagery data and filter coefficients). In one embodiment, the first and the second sets of memory buffers comprise random access memory (RAM). Each of the first and the second sets are logically defined. In other words, respective sizes of the first and the second sets can be reconfigured to accommodate respective amounts of imagery data and filter coefficients.

The first and the second I/O data bus 130a-130b are shown here to connect the CNN processing engines 122a-122h, 132a-132h in a sequential scheme. In another embodiment, the at least one I/O data bus may have different connection scheme to the CNN processing engines to accomplish the same purpose of parallel data input and output for improving performance.

FIG. 2 is a diagram showing an example controller 200 for controlling various operations of at least one CNN processing engine configured on the integrated circuit. Controller 200 comprises circuitry to control imagery data loading control 212, filter coefficients loading control 214, imagery data output control 216, and image processing operations control 218. Controller 200 further includes register files 220 for storing the specific configuration (e.g., number of CNN processing engines, number of input/output data bus, etc.) in the integrated circuit.

Image data loading control 212 controls loading of imagery data to respective CNN processing engines via the corresponding I/O data bus. Filter coefficients loading control 214 controls loading of filter coefficients to respective CNN processing engines via corresponding I/O data bus. Imagery data output control 216 controls output of the imagery data from respective CNN processing engines via corresponding I/O data bus. Image processing operations control 218 controls various operations such as convolutions, rectifications and pooling operations which can be defined by user of the integrated circuit via a set of user defined directives (e.g., file contains a series of operations such as convolution, rectification, pooling, etc.).

More details of a CNN processing engine 302 are shown in FIG. 3. A CNN processing block 304 contains digital circuitry that simultaneously obtains M×M convolutional operations results by performing 3×3 convolutions at M×M pixel locations using imagery data of a (M+2)-pixel by (M+2)-pixel region and corresponding filter coefficients from the respective memory buffers. The (M+2)-pixel by (M+2)-pixel region is formed with the M×M pixel locations as an M-pixel by M-pixel central portion plus a one-pixel border surrounding the central portion. M is a positive integer. In one embodiment, M equals to 14 and therefore, (M+2) equals to 16, M×M equals to 14×14=196, and M/2 equals 7.

FIG. 4 is a diagram showing a diagram representing (M+2)-pixel by (M+2)-pixel region 410 with a central portion of M×M pixel locations 420 used in the CNN processing engine 302.

Imagery data may represent characteristics of a pixel in the input image (e.g., one of the color (e.g., RGB (red, green, blue)) values of the pixel, or distance between pixel and observing location). Generally, the value of the RGB is an integer between 0 and 255. Values of filter coefficients are floating point integer numbers that can be either positive or negative.

In order to achieve faster computations, few computational performance improvement techniques have been used and implemented in the CNN processing block 304. In one embodiment, representation of imagery data uses as few bits as practical (e.g., 5-bit representation). In another embodiment, each filter coefficient is represented as an integer with a radix point. Similarly, the integer representing the filter coefficient uses as few bits as practical (e.g., 12-bit representation). As a result, 3×3 convolutions can then be performed using fixed-point arithmetic for faster computations.

Each 3×3 convolution produces one convolutional operations result, Out(m, n), based on the following formula:

$\begin{array}{cc}\mathrm{Out}\left(m,n\right)=\sum _{1\le i,j\le 3}\mathrm{In}\left(m,n,i,j\right)\times C\left(i,j\right)-b& \left(1\right)\end{array}$ where:

• • m, n are corresponding row and column numbers for identifying which imagery data (pixel) within the (M+2)-pixel by (M+2)-pixel region the convolution is performed;
  • In(m,n,i,j) is a 3-pixel by 3-pixel area centered at pixel location (m, n) within the region;
  • C(i, j) represents one of the nine weight coefficients C(3×3), each corresponds to one of the 3-pixel by 3-pixel area;
  • b represents an offset coefficient; and
  • j are indices of weight coefficients C(i, j).

Each CNN processing block 304 produces M×M convolutional operations results simultaneously and, all CNN processing engines perform simultaneous operations.

FIGS. 5A-5C show three different examples of the M×M pixel locations. The first pixel location 531 shown in FIG. 5A is in the center of a 3-pixel by 3-pixel area within the (M+2)-pixel by (M+2)-pixel region at the upper left corner. The second pixel location 532 shown in FIG. 5B is one pixel data shift to the right of the first pixel location 531. The third pixel location 533 shown in FIG. 5C is a typical example pixel location. M×M pixel locations contains multiple overlapping 3-pixel by 3-pixel areas within the (M+2)-pixel by (M+2)-pixel region.

To perform 3×3 convolutions at each sampling location, an example data arrangement is shown in FIG. 6. Imagery data (i.e., In(3×3)) and filter coefficients (i.e., weight coefficients C(3×3) and an offset coefficient b) are fed into an example CNN 3×3 circuitry 600. After 3×3 convolutions operation in accordance with Formula (1), one output result (i.e., Out(1×1)) is produced. At each sampling location, the imagery data In(3×3) is centered at pixel coordinates (m, n) 605 with eight immediate neighbor pixels 601-604, 606-609.

FIG. 7 is a function diagram showing an example CNN 3×3 circuitry 700 for performing 3×3 convolutions at each pixel location. The circuitry 700 contains at least adder 721, multiplier 722, shifter 723, rectifier 724 and pooling operator 725. In a digital semi-conductor implementation, all of these can be achieved with logic gates and multiplexers, which are generated using well-known methods (e.g., hardware description language such as Verilog, etc.). Adder 721 and multiplier 722 are used for addition and multiplication operations. Shifter 723 is for shifting the output result in accordance with fixed-point arithmetic involved in the 3×3 convolutions. Rectifier 724 is for setting negative output results to zero. Pooling operator 725 is for performing 2×2 pooling operations.

Imagery data are stored in a first set of memory buffers 306, while filter coefficients are stored in a second set of memory buffers 308. Both imagery data and filter coefficients are fed into the CNN block 304 at each clock of the digital integrated circuit. Filter coefficients (i.e., C(3×3) and b) are fed into the CNN processing block 304 directly from the second set of memory buffers 308. However, imagery data are fed into the CNN processing block 304 via a multiplexer MUX 305 from the first set of memory buffers 306. Multiplexer 305 selects imagery data from the first set of memory buffers based on a clock signal (e.g., pulse 312).

Otherwise, multiplexer MUX 305 selects imagery data from a first neighbor CNN processing engine (from the left side of FIG. 3 not shown) through a clock-skew circuit 320.

At the same time, a copy of the imagery data fed into the CNN processing block 304 is sent to a second neighbor CNN processing engine (to the right side of FIG. 3 not shown) via the clock-skew circuit 320. Clock-skew circuit 320 can be achieved with known techniques (e.g., a D flip-flop 322).

The first neighbor CNN processing engine may be referred to as an upstream neighbor CNN processing engine in the loop formed by the clock-skew circuit 320. The second neighbor CNN processing engine may be referred to as a downstream CNN processing engine. In another embodiment, when the data flow direction of the clock-skew circuit is reversed, the first and the second CNN processing engines are also reversed becoming downstream and upstream neighbors, respectively.

After 3×3 convolutions for each group of imagery data are performed for predefined number of filter coefficients, convolutional operations results Out(m, n) are sent to the first set of memory buffers via another multiplex MUX 307 based on another clock signal (e.g., pulse 311). An example clock cycle 310 is drawn for demonstrating the time relationship between pulse 311 and pulse 312. As shown pulse 311 is one clock before pulse 312, results of the 3×3 convolutional operations are stored into the first set of memory buffers after a particular block of imagery data has been processed by all CNN processing engines through the clock-skew circuit 320.

After the convolutional operations result Out(m, n) is obtained from Formula (1), rectification procedure may be performed as directed by image processing control 218. Any convolutional operations result, Out(m, n), less than zero (i.e., negative value) is set to zero. In other words, only positive value of output results are kept. FIG. 8 shows two example outcomes of rectification. A positive output value 10.5 retains as 10.5 while −2.3 becomes 0. Rectification causes non-linearity in the integrated circuits.

If a 2×2 pooling operation is required, the M×M output results are reduced to (M/2)×(M/2). In order to store the (M/2)×(M/2) output results in corresponding locations in the first set of memory buffers, additional bookkeeping techniques are required to track proper memory addresses such that four (M/2)×(M/2) output results can be processed in one CNN processing engine.

To demonstrate a 2×2 pooling operation, FIG. 9A is a diagram graphically showing first example output results of a 2-pixel by 2-pixel block being reduced to a single value 10.5, which is the largest value of the four output results. The technique shown in FIG. 9A is referred to as “max pooling”. When the average value 4.6 of the four output results is used for the single value shown in FIG. 9B, it is referred to as “average pooling”. There are other pooling operations, for example, “mixed max average pooling” which is a combination of “max pooling” and “average pooling”. The main goal of the pooling operation is to reduce the size of the imagery data being processed. FIG. 10 is a diagram illustrating M×M pixel locations, through a 2×2 pooling operation, being reduced to (M/2)×(M/2) locations, which is one fourth of the original size.

An input image generally contains a large amount of imagery data. In order to perform image processing operations. The input image 1100 is partitioned into M-pixel by M-pixel blocks 1111-1112 as shown in FIG. 11A. Imagery data associated with each of these M-pixel by M-pixel blocks is then fed into respective CNN processing engines. At each of the M×M pixel locations in a particular M-pixel by M-pixel block, 3×3 convolutions are simultaneously performed in the corresponding CNN processing block.

Although the invention does not require specific characteristic dimension of an input image, the input image may be required to resize to fit into a predefined characteristic dimension for certain image processing procedures. In an embodiment, a square shape with (2^K×M)-pixel by (2^K×M)-pixel is required. K is a positive integer (e.g., 1, 2, 3, 4, etc.). When M equals 14 and K equals 4, the characteristic dimension is 224. In another embodiment, the input image is a rectangular shape with dimensions of (2¹×M)-pixel and (2¹×M)-pixel, where I and J are positive integers. In yet another embodiment, the image size is 640×640 pixels.

In order to properly perform 3×3 convolutions at pixel locations around the border of a M-pixel by M-pixel block, additional imagery data from neighboring blocks are required. FIG. 11B shows a typical M-pixel by M-pixel block 1120 (bordered with dotted lines) within a (M+2)-pixel by (M+2)-pixel region 1130. The (M+2)-pixel by (M+2)-pixel region is formed by a central portion of M-pixel by M-pixel from the current block, and four edges (i.e., top, right, bottom and left) and four corners (i.e., top-left, top-right, bottom-right and bottom-left) from corresponding neighboring blocks. Additional details are shown in FIG. 12 and corresponding descriptions for the first set of memory buffers.

FIG. 11C shows two example M-pixel by M-pixel blocks 1122-1124 and respective associated (M+2)-pixel by (M+2)-pixel regions 1132-1134. These two example blocks 1122-1124 are located along the perimeter of the input image. The first example M-pixel by M-pixel block 1122 is located at top-left corner, therefore, the first example block 1122 has neighbors for two edges and one corner. Value “0”s are used for the two edges and three corners without neighbors (shown as shaded area) in the associated (M+2)-pixel by (M+2)-pixel region 1132 for forming imagery data. Similarly, the associated (M+2)-pixel by (M+2)-pixel region 1134 of the second example block 1124 requires “0”s be used for the top edge and two top corners. Other blocks along the perimeter of the input image are treated similarly. In other words, for the purpose to perform 3×3 convolutions at each pixel of the input image, a layer of zeros (“0”s) is added outside of the perimeter of the input image. This can be achieved with many well-known techniques. For example, default values of the first set of memory buffers are set to zero. If no imagery data is filled in from the neighboring blocks, those edges and corners would contain zeros.

Furthermore, an input image can contain a large amount of imagery data, which may not be able to be fed into the CNN processing engines in its entirety. Therefore, the first set of memory buffers is configured on the respective CNN processing engines for storing a portion of the imagery data of the input image. The first set of memory buffers contains nine different data buffers graphically illustrated in FIG. 12. Nine buffers are designed to match the (M+2)-pixel by (M+2)-pixel region as follows:

• 1) buffer-0 for storing M×M pixels of imagery data representing the central portion;
• 2) buffer-1 for storing 1×M pixels of imagery data representing the top edge;
• 3) buffer-2 for storing M×1 pixels of imagery data representing the right edge;
• 4) buffer-3 for storing 1×M pixels of imagery data representing the bottom edge;
• 5) buffer-4 for storing M×1 pixels of imagery data representing the left edge;
• 6) buffer-5 for storing 1×1 pixels of imagery data representing the top left corner;
• 7) buffer-6 for storing 1×1 pixels of imagery data representing the top right corner;
• 8) buffer-7 for storing 1×1 pixels of imagery data representing the bottom right corner; and
• 9) buffer-8 for storing 1×1 pixels of imagery data representing the bottom left corner.

Imagery data received from the I/O data bus are in form of M×M pixels of imagery data in consecutive blocks. Each M×M pixels of imagery data is stored into buffer-0 of the current block. The left column of the received M×M pixels of imagery data is stored into buffer-2 of previous block, while the right column of the received M×M pixels of imagery data is stored into buffer-4 of next block. The top and the bottom rows and four corners of the received M×M pixels of imagery data are stored into respective buffers of corresponding blocks based on the geometry of the input image (e.g., FIGS. 11A-11C).

An example second set of memory buffers for storing filter coefficients are shown in FIG. 13A. In one embodiment, a pair of independent buffers Buffer0 1301 and Buffer1 1302 is provided. The pair of independent buffers allow one of the buffers 1301-1302 to receive data from the I/O data bus 1330 while the other one to feed data into a CNN processing block (not shown). Two operational modes are shown herein.

Example storage schemes of filter coefficients are shown in FIG. 13B. Each of the pair of buffers (i.e., Buffer0 1301 or Buffer1 1302) has a width (i.e., word size 1310). In one embodiment, the word size is 120-bit. Accordingly, each of the filter coefficients (i.e., C(3×3) and b) occupies 12-bit in the first example storage scheme 1311. In the second example storage scheme 1312, each filter coefficient occupies 6-bit thereby 20 coefficients are stored in each word. In the third example scheme 1313, 3-bit is used for each coefficient hence four sets of filter coefficients (40 coefficients) are stored. Finally, in the fourth example storage scheme 1314, 80 coefficients are stored in each word, each coefficient occupies 1.5-bit.

In another embodiment, a third memory buffer can be set up for storing entire filter coefficients to avoid I/O delay. In general, the input image must be at certain size such that all filter coefficients can be stored. This can be done by allocating some unused capacity in the first set of memory buffers to accommodate such a third memory buffer. Since all memory buffers are logically defined in RAM (Random-Access Memory), well known techniques may be used for creating the third memory buffer. In other words, the first and the second sets of memory buffers can be adjusted to fit different amounts of imagery data and/or filter coefficients. Furthermore, the total amount of RAM is dependent upon what is required in image processing operations.

When more than one CNN processing engine is configured on the integrated circuit. The CNN processing engine is connected to first and second neighbor CNN processing engines via a clock-skew circuit. For illustration simplicity, only CNN processing block and memory buffers for imagery data are shown. An example clock-skew circuit 1440 for a group of CNN processing engines are shown in FIG. 14. The CNN processing engines connected via the second example clock-skew circuit 1440 to form a loop. In other words, each CNN processing engine sends its own imagery data to a first neighbor and, at the same time, receives a second neighbor's imagery data. Clock-skew circuit 1440 can be achieved with well-known manners. For example, each CNN processing engine is connected with a D flip-flop 1442.

A special case with only two CNN processing engines are connected in a loop, the first neighbor and the second neighbor are the same.

Referring now to FIG. 15, it is a schematic diagram showing an example image processing technique based on convolutional neural networks in accordance with an embodiment of the invention. Based on convolutional neural networks, multi-layer input imagery data 1511a-1511c is processed with convolutions using a first set of filters or weights 1520. Since the imagery data 1511a-1511c is larger than the filters 1520. Each corresponding overlapped sub-region 1515 of the imagery data is processed. After the convolutional results are obtained, activation may be conducted before a first pooling operation 1530. In one embodiment, activation is achieved with rectification performed in a rectified linear unit (ReLU). As a result of the first pooling operation 1530, the imagery data is reduced to a reduced set of imagery data 1531a-1531c. For 2×2 pooling, the reduced set of imagery data is reduced by a factor of 4 from the previous set.

The previous convolution-to-pooling procedure is repeated. The reduced set of imagery data 1531a-1531c is then processed with convolutions using a second set of filters 1540. Similarly, each overlapped sub-region 1535 is processed. Another activation may be conducted before a second pooling operation 1540. The convolution-to-pooling procedures are repeated for several layers and finally connected to at least one Fully-connected (FC) layer 1560. In image classification, respective probabilities of predefined categories can be computed in FC layers 1560.

This repeated convolution-to-pooling procedure is trained using a known dataset or database. For image classification, the dataset contains the predefined categories. A particular set of filters, activation and pooling can be tuned and obtained before use for classifying an imagery data, for example, a specific combination of filter types, number of filters, order of filters, pooling types, and/or when to perform activation. In one embodiment, convolutional neural networks are based on Visual Geometry Group (VGG16) architecture neural nets, which contains 13 convolutional layers and three fully-connected layers.

A trained convolutional neural networks model is achieved with an example set of operations 1600 shown in FIG. 16. At action 1602, a convolutional neural networks model is first obtained by training the convolutional neural networks model based on image classification of a labeled dataset, which contains a sufficiently large number of input data (e.g., imagery data, converted voice data, optical character reorganization (OCR) data, etc.). For example, there are at least 4,000 data for each category. In other words, each data in the labeled dataset is associated with a category to be classified. The convolutional neural networks model includes multiple ordered filter groups (e.g., each filter group corresponds to a convolutional layer in the convolutional neural networks model). Each filter in the multiple ordered filter groups contains a standard 3×3 filter kernel (i.e., nine coefficients in floating point number format (e.g., standard 3×3 filter kernel 1710 in FIG. 17)). Each of the nine coefficients can be any negative or positive real number (i.e., a number with fraction). The initial convolutional neural networks model may be obtained from many different frameworks including, but not limited to, mxnet, Caffe, TensorFlow, etc.

Then, at action 1604, the convolutional neural networks model is modified by converting respective standard 3×3 filter kernels 1710 to corresponding bi-valued 3×3 filter kernels 1720 of a currently-processed filter group in the multiple ordered filter groups based on a set of kernel conversion schemes. In one embodiment, each of the nine coefficients C(i,j) in the corresponding bi-valued 3×3 filter kernel 1720 is assigned a value ‘A’ which equals to the average of absolute coefficient values multiplied by the sign of corresponding coefficients in the standard 3×3 filter kernel 1710 shown in following formula:

$\begin{array}{cc}A=\sum _{1\le i,j\le 3}C\left(i,j\right)\text{/}9& \left(2\right)\end{array}$

Filter groups are converted one at a time in the order defined in the multiple ordered filter groups. In certain situation, two consecutive filter groups are optionally combined such that the training of the convolutional neural networks model is more efficient.

Next, at action 1606, the modified convolutional neural networks model is retrained until a desired convergence criterion is met or achieved. There are a number of well known convergence criteria including, but not limited to, completing a predefined number of retraining operation, converging of accuracy loss due to filter kernel conversion, etc. In one embodiment, all filter groups including those already converted in previous retraining operations can be changed or altered for fine tuning. In another embodiment, the already converted filter groups are frozen or unaltered during the retraining operation of the currently-processed filter group.

Process 1600 moves to decision 1608, it is determined whether there is another unconverted filter group. If ‘yes’, process 1600 moves back to repeat actions 1604-1606 until all filter groups have been converted. Decision 1608 becomes ‘no’ thereafter. At action 1610, coefficients of bi-valued 3×3 filter kernels in all filter groups are transformed from a floating point number format to a fixed point number format to accommodate the data structure required in the CNN based integrated circuit. Furthermore, the fixed point number is implemented as reconfigurable circuits in the CNN based integrated circuit. In one embodiment, the coefficients are implemented using 12-bit fixed point number format.

FIG. 18 is a diagram showing an example data conversion scheme for converting data from 8-bit [0-255] to 5-bit [0-31] per pixel. For example, bits 0-7 becomes 0, bits 8-15 becomes 1, etc.

As described in process 1600 of FIG. 16, a convolutional neural networks model is trained for the CNN based integrated circuit. The entire set of trained coefficients or weights are pre-configured to the CNN based integrated circuit as a feature extractor for a particular data format (e.g., imagery data, voice spectrum, fingerprint, palm-print, optical character recognition, etc.). In general, there are many convolutional layers with many filters in each layer. In one embodiment, VGG16 model contains 13 convolutional layers. In a software based image classification task, computations for the convolutional layers take majority of computations (e.g., 90%) traditionally. The computations is drastically reduced with a dedicated hardware such as CNN based IC 100.

In another embodiment, entire set of coefficients can be trained on the CNN based integrated circuit. In other words, the conversion from full floating point number format to fixed point number format is not necessary. The coefficients of bi-valued 3×3 filter kernels are trained directly. Conversion from standard kernel to bi-value filter kernel is not required.

In one embodiment, fixed point number format is adjustable to fit different requirements (e.g., range of filter coefficients). In other words, number of digits after the radix point may be adjusted.

Referring now to FIG. 19, it is shown a schematic diagram illustrating an example process 1900 of image processing for object detections, image segmentations or optical flow tasks. An input imagery data 1902 is processed through a pixel feature extraction network 1904 to extract fine-grained features in a pixel feature vector. Then pixel feature vector is fed into a task specific network 1906 to obtain output 1908 (e.g., image segmentation, object detection, or optical flow). In the example process 1900, the image processing task (i.e., image segmentation, object detection) is separated into two stages: a general pixel feature extractor 1904 and a task specific network 1906. Each stage is executed and/or trained separately. The general pixel feature extractor 1904 extracts fine-grained features out of the input imagery data 1902. The pixel feature vector is shared by multiple other tasks (e.g., image segmentation, object detection, optical flow, etc.), which is performed in task specific network 1906. Owing to separation of pixel feature extraction network and task specific network, training of these two network models may be performed separately or independently. In particular, the pixel feature extraction network is trained as a generic model, while the task specific network is trained as a specific task such as image segmentation, object detection, or optical flow.

An example pixel feature vector 2020 is created in an example process 2000 shown in FIG. 20. An input imagery data 2002 is processed through a deep learning model 2004 that outputs features with different sizes or sampling locations (i.e., feature maps containing various different scales/resolutions). The different scaled feature maps are combined through skip connections (i.e., short-cuts) to produce a multiple-layer feature map 2010 (e.g., concatenation of several feature maps 2006). Different sizes of feature maps can be matched up by up-sampling, for example, smaller feature map (80×80) is enlarged to match the larger feature map (160×160). The multi-layer feature map 2010 can be optionally compressed to create a pixel feature vector 2020 for task specific network.

An example deep learning network for extracting pixel feature vector based on VGG16 is shown in FIG. 21. Input imagery data 2102 having three channels of 640 pixels by 640 pixels is processed through 13 ordered convolutional layers (Cony. Layers) organized in five groups (i.e., Cony. Layers 1-1, 1-2, Cony. Layers 2-1, 2-2, . . . , Cony. Layers 5-1, 5-2, 5-3). Each group is followed by a 2×2 pooling layer.

In the first group, there are 64 channels of feature maps with each feature map containing a matrix of 640×640 entries. The 640×640 entries correspond to respective pixels in each channel of the input imagery data 2102. Next, feature map contains 320×320 entries with 128 channels in the second group. Then, in the third group, feature map contains 160×160 entries with 256 channels. In the fourth group, feature map includes 80×80 entries with 512 channels. Finally, in the fifth group, feature map contains 40×40 entries with 512 channels. Each entry contains a real number.

In order to create a pixel feature vector 2120 to include fine-grained features from the input imagery data 2102, the last convolutional layer of the third, fourth and fifth groups are combined. Since respective feature maps in the third and fourth groups contain more local details than the one in the fifth group, the combined features in the pixel feature vector 2120 retains higher resolution details from the input imagery data 2102. In order to combine (e.g., concatenation of three feature maps), the feature map from the fourth and the fifth groups must match up with the size of feature map from the third group. In this example, pooling layers are configured for performing 2×2 pooling operations. The feature map from Cony. Layer 5-3 requires 4×4 up-sampling (i.e., converting 40×40 to 160×160), while the feature map from Cony. Layer 4-3 requires 2×2 up-sampling (80×80 to 160×160) to match the size of the feature map from Cony. Layer 3-3. Up-sampling can be done in a number of known techniques, for example, interpolating between two feature values, inserting repeated feature values, etc. As result, pixel feature vector 2120 contains 1280 channels of 160×160 entries. The total number of channels 1280 is the sum of channels in the third, fourth, and fifth groups (i.e., 256, 512, and 512).

An example task specific network 2200 for object detection is shown in FIG. 22. Pixel feature vector (160×160 entries with 1280 channels) 2202 obtained separately is used as an input. Two layers of 3×3 convolution 2204-2206 reduces the feature map to 512 and 128 channels, respectively. The size of the feature map is still 160×160 entries. For the center of an object to be detected, a layer of 3×3 convolution 2212 is followed by a layer of 1×1 convolution 2214. The center 2222 is represented with a feature map having 160×160 feature values. As for the width and height of an object to be detected, a separate layer of 3×3 convolution 2216 is followed by a layer of 1×1 convolution 2218. The resulting width and height 2224 are represented by two feature maps each having 160×160 feature values.

FIG. 23 is a diagram showing an example object detection result. The object's center 2322, width 2324, and height 2326 are the result. FIG. 24 is a diagram showing an example image segmentation result. The background 2410 and an object of interest 2420 are separated by a boundary (bold dotted line).

Although the invention has been described with reference to specific embodiments thereof, these embodiments are merely illustrative, and not restrictive of, the invention. Various modifications or changes to the specifically disclosed example embodiments will be suggested to persons skilled in the art. Whereas the combination of feature maps have been shown and described to concatenate three feature maps, other numbers of feature maps may be used for achieving objectives of the invention, for example, two, four, five. etc. In summary, the scope of the invention should not be restricted to the specific example embodiments disclosed herein, and all modifications that are readily suggested to those of ordinary skill in the art should be included within the spirit and purview of this application and scope of the appended claims.

Claims

1. A pixel feature vector extraction system comprising: a cellular neural networks (CNN) based integrated circuit (IC) configured for extracting a pixel feature vector out of an input imagery data by performing convolution operations using pre-trained filter coefficients of ordered convolutional layers in a deep learning model, the ordered convolutional layers organized in a plurality of groups with each group followed by a pooling layer, said each group configured for a respective feature map, each feature map having a different size, the pixel feature vector containing a combination of feature maps from at least two groups, the combination concatenating respective feature maps of the at least two groups, the at least two groups including a first group and at least one remaining group, wherein the feature map of the first group has the largest size amongst the at least two groups, and wherein a respective size of the feature map of each remaining group is modified to match the size of the feature map of the first group.

2. The pixel feature vector extraction system of claim 1, wherein the CNN based IC comprises a plurality of cellular neural networks (CNN) processing engines operatively coupled to at least one input/output data bus, the plurality of CNN processing engines being connected in a loop with a clock-skew circuit, each CNN processing engine comprising: a CNN processing block configured for simultaneously obtaining results of the convolution operations;a first set of memory buffers operatively coupled to the CNN processing block for storing the input imagery data; anda second set of memory buffers operative coupled to the CNN processing block for storing the pre-trained filter coefficients.

3. The pixel feature vector extraction system of claim 1, wherein the respective feature maps of the at least two groups are from respective last convolutional layers of the at least two groups.

4. The pixel feature vector extraction system of claim 1, wherein the deep learning model comprises Visual Geometry Group's VGG16 model with 13 ordered convolutional layers organized in 5 groups.

5. The pixel feature vector extraction system of claim 1, wherein each of the pre-trained filter coefficients comprises bi-valued 3×3 filter kernel.

6. The pixel feature vector extraction system of claim 5, wherein the pre-trained filter coefficients are obtained as a generic model.

7. The pixel feature vector extraction system of claim 1, wherein the CNN based IC is further configured for performing activation.

8. The pixel feature vector extraction system of claim 1, further comprises another processor for performing task specific neural network.

9. The pixel feature vector extraction system of claim 8, wherein the task specific neural network contains at least one 3×3 convolutional layer and at least one 1×1 convolutional layer.

10. The pixel feature vector extraction system of claim 8, wherein the task specific neural network is configured for object detection.

11. The pixel feature vector extraction system of claim 8, wherein the task specific neural network is configured for image segmentation.

12. The pixel feature vector extraction system of claim 8, wherein the task specific neural network is configured for optical flow application.