KR100244855B1 - Bit control method of region encoding - Google Patents

Abstract

The present invention relates to a region-coded bit number control method for a differential encoder, in which image information on a region of interest and other regions is encoded with a different number of bits according to the region-specific interest, and thus can be accurately transmitted to a transmission path.

To this end, according to the present invention, a method for controlling the number of bits per region for a differential encoder according to the present invention includes separating the image information into a plurality of regions including pixels having the same brightness, and the image information of the region to be currently encoded is the image of the region of interest. A second step of determining whether the information is information; a third step of detecting the size of the remaining storage area of the buffer connected to the transmission path; whether image information of the area to be currently encoded is the region of interest and the remaining storage area of the buffer. And a fourth step of varying the quantization parameter for the region to be currently encoded according to the size.

Description

Method of Controlling the Number of Coded Bits by Region for Differential Encoders

1 is a block diagram of a differential encoder to which the method for controlling the number of encoded bits per region according to an embodiment of the present invention is applied.

2 is a flowchart illustrating a method for controlling the number of encoded bits for each region of a differential encoder according to an embodiment of the present invention.

3 is a diagram showing the state of the buffer shown in FIG.

<Explanation of symbols for main parts of drawing>

10: segmentation unit 12, 14: first and second coding unit

16: buffer 18: region of interest candidate extractor

20: recognition unit 22: buffer controller

24: database

The present invention relates to a differential encoder for differentially encoding image information to be transmitted through a transmission path for each region, and more particularly, to a method for controlling the number of encoded bits for a region for a differential encoder for controlling the image information to be encoded with a different number of bits for each region.

In general, the amount of image information becomes larger as the image quality becomes higher. This large amount of video information is difficult to transmit via a limited bandwidth transmission path and is compressed before being transmitted. Various methods are used for a compression algorithm for compressing image information, but a differential encoder implemented by a differential encoding method having the highest compression ratio of image information is commonly used. The differential encoder divides an image signal into a plurality of areas including pixels having the same brightness information, and encodes color information, contour information, and motion information of each area. The differential encoder transmits the encoded image information to the transmission path via a buffer connected to the transmission path.

However, the conventional differential coder encodes the color information and the contour information for each region by the same number of bits irrespective of the importance of the region, thereby encoding portions other than the region of interest with the same image quality as the region of interest. As a result, the image information of the ROI transmitted to the transmission path by the conventional differential encoder is virtually impossible to be restored to the high picture quality at the reception side. In addition, the buffer that relays the differential encoder and the transmission path cannot accurately transmit the image information of the ROI having a large amount of data to the transmission path when the image information is encoded with a different number of bits depending on whether the image information is the ROI and other areas. This is because image information of a region of interest containing a large amount of data overflows from the buffer.

Accordingly, an object of the present invention is to provide a method of controlling the number of encoded bits per region for a differential encoder, in which image information about a region of interest and other regions is encoded with a different number of bits according to the region-specific interest, and thus can be accurately transmitted to a transmission path. Is in.

In order to achieve the above object, the method of controlling the number of bits per region for a differential encoder of the present invention includes a first step of dividing image information into a plurality of regions including pixels having the same brightness, and image information of a region to be currently encoded. A second step of determining whether the area is image information, a third step of detecting a size of the remaining storage area of the buffer connected to the transmission path, whether the image information of the area to be currently encoded is a region of interest and the remaining of the buffer And a fourth step of varying the quantization parameter for the region to be currently encoded according to the size of the storage region.

Other objects and advantages of the present invention other than the above objects will become apparent from the following description of the preferred embodiments with reference to the accompanying drawings.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to FIGS. 1 to 3.

1 is a differential encoder to which the method for controlling the number of encoded bits per region according to an exemplary embodiment of the present invention is applied. In FIG. 1, a differential encoder includes a segmentation unit 10 for inputting a video signal from an input line 11, and first and second codings commonly connected to the segmentation unit 10. In FIG. And a buffer 16 for transmitting the coded signals from the coding units 12 and 14 to the transmission path 19. The segmentation unit 10 divides the image signal from the input line 11 into a plurality of areas including pixels having the same color information to generate contour information, color information, motion information, and the like for the areas. This color information, motion information and contour information are supplied to the first and second coding units 12 and 14 via the first node 13. The first coding unit 12 encodes the color information, the motion information, and the contour information of the background among the color information, the motion information, and the contour information from the first node 13 with a relatively small number of bits. Meanwhile, the second coding unit 14 encodes the color information, the motion information, and the contour information of the region of interest among the color information, the motion information, and the contour information from the first node 13 with a relatively large number of bits. These first and second coding units 12 and 14 are selectively driven depending on whether the color information, motion information and contour information from the first node 13 are in the ROI or other regions. That is, when the color information, motion information, and the contour information from the first node 13 are not related to the ROI, and the second coding unit 14 determines the first node 13. The color information, the motion information and the contour information from the data are driven when they relate to the ROI. As a result, the second coding unit 14 allocates a large number of bits to the color information, the motion information, and the contour information of the ROI, so that the image information of the ROI can be maintained in a higher quality than the background region. The buffer 16 temporarily stores the encoded image information from the first and second coding units 12 and 14 and then transmits the encoded image information toward the transmission path 19.

The differential encoder further includes a region of interest candidate extractor 18, a recognizer 20, and a buffer controller 22 connected in series with the first node 13. The ROI candidate extractor 18 extracts ROI candidates that can be treated as ROI according to the change state of color information, motion information, and contour information for each region from the first node 13. The ROI candidate extractor 18 supplies color information, motion information, and contour information of the ROI candidates to the recognition unit 20. The recognition unit 20 searches color information, motion information, and contour information of candidate regions of interest according to reference data to determine whether image information of a region currently being input is a region of interest. The reference data used to determine the ROI is supplied from the database 24 indicated by a dotted line to the recognition unit 20, and includes information on importance of each ROI when there are ROIs of various importance. The recognition unit 20 generates an ROI determination signal having a different logical state depending on whether image information of an area currently being input is an ROI. The ROI decision signal is supplied to the first and second coding units 12 and 14 and the buffer controller 22 via the second node 15. The first and second coding units 12 and 14 are complementarily driven by the ROI decision signal. In addition, the recognition unit 20 supplies importance information on the image information of the area currently being input to the buffer controller 22 together with the ROI determination signal. The buffer controller 22 determines the target predicted bit number of the region to be encoded by the ROI determination signal and the importance information. The buffer controller 22 adjusts the number of bits (that is, the quantization parameter) to be encoded by the first and second coding units 12 and 14 according to the current state of the buffer 16. This quantization parameter is supplied from the buffer controller 22 to the first and second coding units 12 and 14 via the third node 17. Then, the first and second coding units 12 and 14 encode the image information of the current input area into the number of bits corresponding to this quantization parameter. As a result, the buffer 16 is optimally controlled by the buffer controller 22, and the image information of the region of interest is restored to a high quality state by allocating the number of bits differently depending on whether the region-specific image information is applied to the region of interest. To be encoded.

2 is a flowchart illustrating a method of controlling the number of coded bits applied to the differential encoder shown in FIG. Before describing this flowchart, let's define some terms.

b _{pp_min} (r _i ): Minimum number of encoded bits per pixel included in r _i area

b _{pp_obi} (r _i ): The number of target coded bits per pixel included in the r _i region

Rs: auxiliary information such as frame start information, motion information, split information

a: inter-block coding block (inter block; block in case of motion compensation between previous block and current block)

b: intra-block (intra-block) A block in which only current block data is encoded and transmitted.

영역에 대한 최소 허용화질을 만족시키는데 필요한 전체 비트 수

Total number of bits needed to meet minimum acceptable picture quality for region

영역에 대한 목표 허용화질을 만족시키는데 필요한 전체 비트 수

Total number of bits needed to meet target tolerance for region

R _γi : Number of bits used to encode r _i region

R _t = R / frame rate: target bits per frame

BF_init: Initial buffer state

: 이전 프레임까지 부호화한 경우의 버퍼상태

: Buffer status when encoding up to previous frame

If the bit rate per second is defined as R, R _{cur_frame, which} is the number of prediction frame bits for the current video, is determined as follows.

R _{cur_frame} = (BF_init-BF_pre) + R _t ------------------ (Equation 1)

That is, the number of prediction bits available per current frame is a value obtained by subtracting the state of the buffer from the initial buffer state to the previous frame plus the target frame rate. Therefore, the number of bits available for encoding the DCT coefficient of each block data in the actual current frame is represented by the following equation.

R _{cur_frame_dct} = R _t + (BF_init-BF_pre)-Rs ----------- (Equation 2)

Next, the flowchart shown in FIG. 2 will be described in conjunction with the differential encoder shown in FIG. First, in a thirtieth step, the buffer controller 22 determines the number of bits (R _{cur_frame_dct} ) that can be used to encode DCT coefficients of the area data in the current frame by the above equation (2). In addition, the buffer controller 22 inputs a region of interest determination signal indicating whether the region to be currently encoded is the region of interest from the recognizer 20 (step 32), and the importance information about the region to be currently encoded r _i . M _r ) is input from the recognition unit 20 (step 34). Here, the importance information M _r is expressed as M _{γ 1} > M _{γ 2} . 〉 M _γi 〉 It is said that importance information having a value of M _{γ i} is input among M _{γ n} . In addition, the buffer controller 22 calculates r _i region can target prediction bits for (bpp _obj (r ^a _i) _cur) using a region of interest determination signal and the priority information (M _γi) by the following equation (3) (Step 36).

Here, (_) _ cur represents the area (·) of the current frame, and (·) _pre represents the area (·) of the previous frame.

The buffer controller 22 determines the state in which data is stored in the buffer 16 (that is, the state of the current buffer) (steps 38 to 42). In detail, the buffer controller 22 determines that the number of bits Ri_remain of the remaining storage area in which data can be stored in the buffer 16 as shown in FIG. 3 satisfies the minimum allowable quality min_sp_r _i. ) And a value between the number of bits needed to satisfy the target tolerance (max_sp_r _i ) (step 38), a value larger than the number of bits required to satisfy the target tolerance (max_sp_r _i ) (step 40), or the minimum acceptable quality. It is determined whether it has a value (step 42) smaller than the number of bits min_sp_r _i required to satisfy. When the number of bits Ri_remain of the remaining storage area is greater than the number of bits max_sp_r _i required to satisfy the target permissible quality in step 40, that is, the number of bits Ri_remain of the remaining storage area is greater than or equal to the max level in FIG. In the case of having a size, the buffer controller 22 determines that the number of bits of the storage area remains on the buffer 16 even after allocating the number of bits per pixel to satisfy the target permissible picture quality for the region r _{i to} be currently encoded. At this time, the buffer controller 22 increases the number of encoded bits per pixel by lowering the value of the quantization parameter QP for the region r _i to be currently encoded to be supplied to the first and second coding units 12 and 14. In this case, the image information of the current region is maintained in high quality (step 44). This quantization parameter QP is calculated by the following expression (4).

k △x ------------------------------(식4)△ QP

k △ x ------------------------------ (Equation 4)

In Equation 4, k is a constant and Δx represents an extra bit. However, if it is determined that the encoded information on the region (r _i ) to be currently encoded is in the underflow altert state when it is stored in the buffer 16, that is, the bit is allocated to satisfy the target permissible picture quality. When the state of the buffer 16 is BF_un_alert as shown below (Equation 5), that is, when it is determined that the number of bits of the remaining storage area of the buffer 16 has a size less than or equal to the min level in FIG. ) Inserts a stuffing bit to get out of the BF_un_alert state.

In Equation 5, BF_un_alert means the depletion warning level of the buffer 16.

If the remaining storage area Ri_remain is smaller than the number of bits min_sp_r _i that satisfies the minimum allowable quality in step 42, that is, the number of bits of the remaining storage area of the current buffer 16 in FIG. In the case of having a size, encoding of the image information of the region r _{i to} be currently encoded is omitted.

On the other hand, when having a size in the range 38 step residual storage (Ri_remain) the number of bits to meet a minimum acceptable image quality (min_sp_r _i) and to satisfy the target allows the image quality of Bits (max_sp_r _i) is required, the buffer controller (22 ) is (the number of target prediction bits for the r _i) (bpp _obj (r ^a _i), the current area to be encoded _cur) is "the number of bits in the remaining storage areas (Ri_remain) - number of bits to meet a minimum acceptable image quality (min_sp_r _i) The value between + 1 "and the" number of bits of remaining storage (Ri_remain)-number of bits required to satisfy the target tolerance (max_sp_r _i ) + 1 "," the number of bits of remaining storage (Ri_remain)-target " A value larger than the number of bits required to satisfy the allowable quality (max_sp_r _i + 1 ") or a value smaller than the" number of bits in the remaining storage area (Ri_remain)-the number of bits satisfying the minimum allowable quality (min_sp_r _i ) + 1 ". The eggplant is examined (steps 46 to 50).

In step 46, the target prediction bit number bpp _obj (r ^a _i ) _cur for the region r _{i to} be currently encoded is “the number of bits of the remaining storage region Ri_remain—the number of bits satisfying the minimum allowable image quality min_sp_r _i ) + 1 " and a value between " number of bits of remaining storage Ri_remain-number of bits required to satisfy the target permissible quality max_sp_r _i + 1 &quot;, the buffer controller 22 requests A quantization parameter QP capable of satisfying the obtained image quality is calculated (step 52). However, the value of the quantization parameter (QP) is determined from the information on a given area in advance. If the information on a given area is not known accurately, only the information on the given area is determined. Create a database from this set and use it as the initial value. If the characteristics of the given area and the corresponding quantization parameter (QP) are identified in the process, the previous quantization parameter (QP) is substituted for these values.

In step 48, the target prediction bit number bpp _obj (r ^a _i ) _cur for the region r _{i to} be currently encoded is “the number of bits of the remaining storage region Ri_remain—the number of bits necessary to satisfy the target allowable quality ( When the value is larger than max_sp_r _i ) + 1 ", the buffer controller 22 determines that an extra bit remains even after bit allocation is performed with the target bit in the region r _{i to} be encoded. At this time, the buffer controller 22 lowers the value of the quantization parameter QP by calculating the value of the quantization parameter QP to be supplied to the first and second coding units 12 and 14 by Equation 5 above. (Step 54). As a result, the image information of the region r _{i to} be currently encoded is encoded with a large number of bits per pixel to maintain high quality.

In step 50, the target predicted bit number bpp _obj (r ^a _i ) _cur for the region r _{i to} be currently encoded is “the number of bits Ri of the remaining storage region Ri_remain—the number of bits satisfying the minimum allowable image quality min_sp_r _i ) If the value is less than + 1 ", the buffer controller 22 determines that there is not enough storage space left for the region r _{i to} be currently encoded to allocate bits to the target bits. At this time, the buffer controller 22 increases the value of the quantization parameter QP for the region r _{i to} be currently encoded, and thus estimates the required amount from the relationship between the number of remaining bits that can be stored on the buffer 16 and the quantization parameter QP. Do not exceed the number of bits (step 56).

Most of the image information encoded by the first and second coding units 12 and 14 according to the quantization parameter thus determined does not cause flooding of the buffer 16. That is, the image information encoded according to the quantization parameter calculated in step 54 satisfies the following expression (6).

Here, BF_ov_altert means the flood warning level of the buffer 16.

The buffer controller 22 supplies the first and second coding units 12 and 14 to the first and second coding units 12 and 14 by supplying the quantization parameter QP determined in steps 44, 52, 54, or 56. In operation 58, the image information of the region r _i to be encoded is encoded. Subsequently, the buffer controller 22 checks whether the region count value i in the self has reached a threshold and determines whether image information of all regions has been encoded (step 60). If there is still an area to be encoded in step 60, that is, the area count value i has not reached the threshold, the buffer controller 22 adds the area count value i by "1" and then Return to step 38 (step 62). On the other hand, when the area count value i reaches the threshold in step 60, the buffer controller 22 ends the encoding operation.

As described above, in the method for controlling the number of bits for each region of the differential encoder according to the embodiment of the present invention, the image information of the region of interest is changed by varying the number of encoding bits of the image information for each region according to the interest of each region and the state of the buffer. It is possible to keep the information high quality and to prevent the overflow of the buffer. Accordingly, the method of controlling the number of bits for each region of the differential encoder according to the embodiment of the present invention provides an advantage of encoding the image information of the ROI in a high quality state and accurately transmitting the encoded image information to the transmission path.

As described above, the present invention has been described with reference to the embodiment illustrated in FIG. 2, but those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention. Therefore, the technical scope of the present invention should not be limited to the contents described in the detailed description of the specification but should be defined by the claims.

Claims (4)

A first step of dividing the image information into a plurality of regions including pixels of the same brightness, a second step of determining whether the image information of the region to be currently encoded is the image information of the region of interest, a buffer connected to the transmission path Detecting a size of the remaining storage area of the variable; varying the quantization parameter for the current area to be encoded according to whether image information of the current area to be encoded is a region of interest and the size of the remaining storage area of the buffer; And a fourth step of controlling the number of encoded bits per region for a differential encoder. The method of claim 1, wherein the third step includes a value between the number of bits for the remaining number of bits in the remaining storage area of the buffer and the number of bits required to satisfy a target allowable quality, and the bits necessary for satisfying a target allowable quality. And determining one of the values larger than the number and smaller than the number of bits satisfying the minimum permissible picture quality. The method of claim 2, further comprising a fifth step of omitting encoding of image information of a region to be currently encoded when the number of bits of the remaining storage region of the buffer is smaller than the number of bits satisfying a minimum allowable picture quality. A method for controlling the number of encoded bits per region for a differential encoder. 3. The method of claim 2, further comprising determining a target expected number of bits for the region to be currently encoded using the degree of interest when the image information of the region to be currently encoded is image information of the region of interest. And a method of controlling the number of encoded bits per region for a differential encoder.

Images