Patent Application
20080175473
This application claims priority from Korean Patent Application No. 10-2007-0006293, filed on Jan. 19, 2007, in the Korean Intellectual Property Office, the disclosure of which is incorporated herein in its entirety by reference.
1. Field of the Invention
The present invention relates to a device and a method of estimating noise of an image signal and an apparatus and a method of converting an image signal using the same, and more particularly, to a method and a device for effectively removing noise from an image by accurately estimating the noise of a frame based image.
2. Description of the Related Art
In general, an image such as a film for a theater movie is produced as an interlaced field image by interlacing a progressive frame image. A 3:2 pull-down format technique, or a 2:3:3:2 cadence pattern technique is employed as a method of interlacing an image. A field image obtained by the aforementioned method is referred to as a frame based field image.
When the interlaced field image is deinterlaced and recovered as the progressive frame image, the original image is interpolated using temporal interpolation that is a simple intra-interpolation technique. However, when the interlaced field image is interpolated through the temporal interpolation technique, the recovered image deteriorates due to jitter, judder, and other noise. Accordingly, in order to reduce the deterioration of the recovered image, additional processing is needed.
The present invention provides provide a device and a method of estimating noise of an image signal so as to effectively cancel an influence of the noise during interpolation of a frame based field image and an apparatus and a method using the same.
According to an aspect of the present invention, there is provided a device for estimating noise of an image, the device including a frame sum of absolute difference (SAD) calculation unit which calculates a plurality of frame SADs by using a plurality of image fields; a minimum SAD detection unit which determines a minimum frame SAD among the plurality of frame SADs calculated by the frame SAD calculation unit; and a noise level determination unit which determines noise levels of the plurality of image fields by using the minimum frame SAD detected by the minimum frame SAD detection unit.
According to another aspect of the present invention, there is provided an apparatus for converting an image signal, the apparatus including a noise estimation device which determines a noise level by using a minimum frame SAD detected from a plurality of input image fields; an interpolation filter which interpolates the input image fields; and an interpolation filter control unit which generates a filter coefficient control signal of the interpolation filter by using the noise level obtained by the noise estimation device.
According to yet another aspect of the present invention, there is provided a method of estimating noise of an image signal, the method including calculating a plurality of frame SADs by using a plurality of image fields; determining the minimum frame SAD among the plurality of frame SADs; and determining the noise level by using the minimum frame SAD.
According to yet another aspect of the present invention, there is provided a method of converting an image signal, the method including estimating noise by using a minimum frame SAD detected by using a plurality of input image fields; and controlling an interpolation filter of the input image fields by using the estimation result of the noise.
The above and other aspects of the present invention will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings in which:
First, backgrounds of exemplary embodiments of the present invention will be briefly described.
An apparatus and a method of converting an image according to an exemplary embodiment of the present invention to be described are based on a frame based interlaced field. That is, an exemplary embodiment of the present invention is based on an interlaced field obtained by interlacing theater movie film using a 3:2 pull-down format technique or 2:3:3:2 cadence pattern technique, etc. Such frame based field should be distinguished from general TV's interlaced field.
Referring to
Now, exemplary embodiments of the present invention will be described in detail with reference to the attached drawings.
An apparatus for converting an image signal includes an image noise estimation device 21, an interpolation filter control unit 22, and an interpolation filter 23. The image noise estimation device 21 includes a frame SAD calculation unit 211, a minimum frame SAD detection unit 212, a first register 214, and a noise level determination unit 213.
The frame SAD calculation unit 211 includes field delays (FDs) 31 and 32, an absolute difference (AD) calculation unit 33, and an AD summation unit 34.
An image signal of a first field (e.g., A-top field 111) among a sequence of image fields for obtaining the frame SAD is input into the frame SAD calculation unit 211. The input image signal is field-delayed twice through the first and second FDs 31 and 32 and input into the AD calculation unit 33.
At the same time, the AD calculation unit receives the image signal of the second field (e.g., A-top field 113) and calculates an AD of pixel intensity for each pixel or each block.
The AD summation unit 34 sums up the ADs for each field and outputs the sum of the ADs. The sum is referred to as frame SAD (sum of absolute difference). It has to be noted that the pair of fields for obtaining the frame SAD is selected so that the pair of fields has the same phase (top or bottom) (for example, 111-113, 112-121, 113-122, and the like).
The minimum frame SAD detection unit 212 includes (N−1)th number of field delay units 41 to 44 and a minimum SAD selection unit 45.
N number of frame SADs which are output from the frame SAD calculation unit 211 are sequentially input into the minimum SAD detection unit 212 and input into the minimum SAD selection unit 45 via the field delay units 0 to N−1. The minimum SAD selection unit selects and outputs the minimum SAD by comparing the magnitudes of the N number of frame SADs. The number N may be selected to be the number of frame SADs which constitutes a period (for example, N=5, in
The noise level determination unit 213 may include a storage unit, which stores a threshold value and pattern matching information, and a comparison unit.
The noise level determination unit 213 of
The noise level determination unit 213 of
The interpolation filter 23 in
An image signal 24 of
In operation 61, frame SADs are calculated with a plurality of pairs of fields. An operation of selecting two nearest fields having the same phase (top or bottom) and calculating a frame SAD is repeated N times. When the A-top field 111 is selected, the nearest top field is another A-top field 113, and therefore the A-top field 113 is selected. The frame SAD (FSAD1) between the two fields 111 and 113 is obtained. Then, A-bottom field 112 and B-bottom field 121, which is the nearest field having the same phase, are selected and a frame SAD (FSAD2) is obtained using the two fields. A series of frame SADs calculated by repeating the aforementioned method are illustrated in graphs in
The minimum frame SAD is detected in operation 62. As shown in
In operation 63, a noise level of an image signal is determined by using the minimum frame SAD. As described above, since the minimum frame SADs (e.g., FSAD1 and FSAD2) are calculated from the pair of image fields having the same phase which are extracted from the same image frame, the minimum frame SAD is minimum (ideally zero). When the minimum frame SAD is not zero, the minimum frame SAD corresponds to the noise of the image field. Accordingly, as the difference between the minimum frame SAD and the predetermined threshold value increases, it is determined that the noise level of the image signal increases.
In operation 64, the interpolation filter 23 is controlled by using the noise level.
As described above, the interpolation filter 23 according to an exemplary embodiment of the present invention includes not only a function of interpolation of an input image signal but also functions of jitter detection, judder detection, detail enhancement, and the like so as to remove noise such as jitter, judder, and the like from the input image signal. Accordingly, the function and the performance of the interpolation filter 23 can be changed by adjusting various coefficients of the interpolation filter 23. Accordingly, interpolation filters can be variously controlled based on the noise level obtained from the noise level determination unit. For example, when the noise level is high, the filter coefficient is adjusted so that cadence interpolation which interpolates the original image is weakened, while jitter detection and pixel judder detection is strengthened. In addition, when the noise level is high, the filter coefficient is adjusted so that the detail enhancement sensitive to the noise is weakened.
In addition, it is determined whether noise level increase is temporary by using a noise level history in which the noise levels are recorded for a predetermined period. If the noise level temporarily increases, the noise is ignored and the filter coefficient is not adjusted.
An exemplary embodiment of the present invention provides a method of estimating noise using the minimum frame SAD. Accordingly, it is possible to effectively and accurately estimate noise of a frame based image field.
In addition, noise such as judder, jitter and the like of an image signal is effectively removed by controlling interpolation of the image signal by using the noise level obtained by estimating the noise.
Exemplary embodiments of the present invention can also be embodied as computer readable codes on a computer readable recording medium. The computer readable recording medium is any data storage device that can store data which can be thereafter read by a computer system. Examples of the computer readable recording medium include read-only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tapes, floppy disks, optical data storage devices, and carrier waves (such as data transmission through the Internet). The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.
While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. The exemplary embodiments should be considered in descriptive sense only and not for purposes of limitation. Therefore, the scope of the invention is defined not by the detailed description of the invention but by the appended claims, and all differences within the scope will be construed as being included in exemplary embodiments of the present invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10-2007-0006293 | Jan 2007 | KR | national |
