KR102163556B1 - Method and Apparatus for Detecting Scene Change of Nighttime Image in Compression Domain - Google Patents

Abstract

In the present embodiments, a ratio of an intra prediction coding unit among all prediction coding units is calculated for each frame, a scene change candidate is selected based on the ratio of the intra prediction coding unit, and the number of prediction coding units encoded in the skip mode is total By excluding from the number of prediction coding units and detecting a scene change using the complexity and edge histogram calculated based on the intra prediction coding unit, the actual scene change is finally achieved through a scene change evaluation criterion that is robust to the brightness of the image in the compressed region. It provides a method and apparatus for detecting a scene change of a night image capable of detecting a frame close to a frame.

Description

[Method and Apparatus for Detecting Scene Change of Nighttime Image in Compression Domain}

The technical field to which the present invention pertains relates to a method and apparatus for detecting a scene change in a compressed image.

The content described in this section merely provides background information on the present embodiment and does not constitute the prior art.

With the development of network transmission technology and video encoding technology, the demand for real-time high-definition large-capacity video content is exploding. Since the video has a relatively large playback time and file size, it is difficult for a viewer to quickly grasp the overall contents of the video during a given time. When a key frame containing key information of a video is detected in the video, a viewer can roughly grasp the overall content of the video relatively quickly even with only the detected key frame.

In general, a frame in which a rapid transition between scenes of a video is made in time is determined as a key frame. The new scene information is determined based on the fact that there is no correlation between the frame in which the scene change occurred and the adjacent previous frame. Existing scene change detection techniques are largely divided into a scene change detection method in a spatial domain and a scene change detection method in a compressed area.

The method of using the edge histogram in the spatial domain is to have the total number of histogram bins per frame (number of sub-images * number of edge directions), create a histogram with the edge distribution as bins, and then use the edge histogram difference between adjacent frames. Determine whether to change the scene. In this method, since gray level values of actual pixels per frame are required to construct an edge histogram in the spatial domain, a compressed image must be decoded. In addition, in order to detect the edge distribution, there is a problem in that the sub-picture is divided into k video blocks, the video block is divided into n sub-blocks, and a filter coefficient is applied to the average gray level value of the sub-blocks to undergo a filtering process. .

The method of detecting scene change using intra-prediction in the compressed region is based on the tendency that the detailed part of the image is encoded as a relatively small intra-prediction block. I judge that is high. A scene change is determined based on a distribution change of a block encoded in an intra prediction mode for each frame. Since this method considers frames in which only blocks encoded in the intra prediction mode exist, the detection method when there is no frame encoded in the intra prediction mode or blocks encoded in the inter prediction mode are distributed together. There is a problem that cannot be used.

Korean Registered Patent Publication No. 10-0816013 (2008.03.17.)

Embodiments of the present invention calculate the ratio of intra-mode encoded blocks by excluding the number of blocks encoded in the skip mode for a night image or a low-light image at night from the total number of blocks, so that a frame is sufficiently qualified as a scene change candidate. Edo has a main purpose to solve the problem of not being detected by skip mode blocks due to low brightness.

Still other objects, not specified, of the present invention may be additionally considered within the range that can be easily deduced from the following detailed description and effects thereof.

According to an aspect of the present embodiment, in a method for detecting a scene change of a night image by a computing device, a ratio of a prediction coding unit in a compression region encoded in an intra mode from a frame of an image captured at night is calculated, and the intra Selecting a scene change candidate using the ratio of the prediction coding units in the compression region encoded in the mode, and extracting from the compression region based on the intra mode of the prediction coding unit in the compression region encoded in the intra mode It provides a method for detecting a scene change of a night image including the step of detecting a scene change using a complexity and an edge histogram.

According to another aspect of the present embodiment, a ratio of prediction coding units in a compression region encoded in an intra mode from a frame of an image photographed at night is calculated, and a ratio of prediction coding units in a compression region encoded in the intra mode is calculated. A candidate selection unit that selects a scene change candidate using, and in the compression region, detects a scene change using a complexity and edge histogram extracted based on an intra mode of a prediction coding unit in the compression region encoded in the intra mode. It provides an apparatus for detecting a scene change of a night image including a scene determination unit.

According to another aspect of the present embodiment, a computer program for detecting a scene change of a night image by being recorded on a non-transitory computer-readable medium including computer program instructions executable by a processor, the computer program instruction When they are executed by at least one processor of the computing device, the ratio of prediction coding units in the compression region encoded in the intra mode from the frame of the image captured at night is calculated, and in the compression region encoded in the intra mode Selecting a scene change candidate using the ratio of the prediction coding units of, and in the compression region, using a complexity and edge histogram extracted based on the intra mode of the prediction coding unit in the compression region encoded in the intra mode. There is provided a computer program that performs operations including detecting a scene change.

As described above, according to embodiments of the present invention, a ratio of an intra prediction coding unit among all prediction coding units for each frame is calculated, a scene change candidate is selected based on the ratio of the intra prediction coding unit, and a skip mode The number of prediction coding units encoded by is excluded from the total number of prediction coding units, and a scene change is detected using a complexity and an edge histogram based on an intra prediction mode in the compression region, thereby using H.264 or a high definition video compression standard. It is not very dependent on the syntax of the HEVC standard, and can be implemented flexibly within the existing standard with relatively simple syntax independent, and through the scene change evaluation standard that is robust to the brightness of the image without complete decoding of the image in the compressed area There is an effect of being able to detect adjacent frames.

Even if it is an effect not explicitly mentioned herein, the effect described in the following specification expected by the technical features of the present invention and the provisional effect thereof are treated as described in the specification of the present invention.

1 is a diagram illustrating an image encoding apparatus.
2 is a diagram illustrating an image decoding apparatus.
3 is a block diagram illustrating a scene change detection apparatus according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating that a scene change detection apparatus according to an embodiment of the present invention divides into a PU unit according to an HEVC prediction mode.
5 is a diagram illustrating an edge direction according to an intra prediction mode referenced by a scene change detection apparatus according to an embodiment of the present invention.
6 is a diagram illustrating that a scene change detection apparatus according to an embodiment of the present invention constructs a complexity and edge histogram from one frame.
7 is a diagram illustrating that a scene change detection apparatus according to an embodiment of the present invention determines a prediction mode by rate-distortion optimization.
8 is a flowchart illustrating a method for detecting a scene change according to another embodiment of the present invention.
9 is a diagram illustrating a macroblock mode distribution of a scene among night black box captured images processed by the scene change detection apparatus according to an embodiment of the present invention.

Hereinafter, in describing the present invention, when it is determined that the subject matter of the present invention may be unnecessarily obscured as matters apparent to those skilled in the art with respect to known functions related to the present invention, a detailed description thereof will be omitted, and some embodiments of the present invention will be described. It will be described in detail through exemplary drawings.

1 is a diagram illustrating an image encoding apparatus.

Referring to FIG. 1, the image encoding apparatus 100 includes a motion prediction unit 111, a motion compensation unit 112, an intra prediction unit 120, a switch 115, a subtractor 125, and a transform unit 130. , A quantization unit 140, an entropy encoding unit 150, an inverse quantization unit 160, an inverse transform unit 170, an adder 175, a filter unit 180, and a reference image buffer 190.

The image encoding apparatus 100 encodes an input image in an intra mode or an inter mode, and outputs a bitstream. Hereinafter, in an embodiment of the present invention, intra prediction may have the same meaning as intra prediction, and inter prediction may have the same meaning as inter prediction. In order to determine an optimal prediction method for the prediction unit, an intra prediction method and an inter prediction method may be selectively used for the prediction unit. The image encoding apparatus 100 generates a prediction block for an original block of an input image, and then encodes a difference between the original block and the prediction block.

In the case of the intra prediction mode, the intra prediction unit 120 (or the intra prediction unit can also be used as a term having the same meaning) performs spatial prediction using pixel values of an already coded block around the current block. Create a block.

In the case of the inter prediction mode, the motion prediction unit 111 finds a region in the reference image stored in the reference image buffer 190 that best matches the input block in the motion prediction process to obtain a motion vector. The motion compensation unit 112 generates a prediction block by performing motion compensation using a motion vector.

The subtractor 125 generates a residual block based on the difference between the input block and the generated prediction block. The transform unit 130 outputs a transform coefficient by performing transform on the residual block. Further, the quantization unit 140 quantizes the input transform coefficient according to the quantization parameter and outputs a quantized coefficient. The entropy encoding unit 150 entropy-encodes the input quantized coefficients according to a probability distribution and outputs a bit stream.

HEVC also performs intra prediction encoding (intra prediction encoding) or inter prediction encoding (inter prediction encoding). During inter prediction encoding, the currently encoded image needs to be decoded and stored in order to be used as a reference image. Accordingly, the quantized coefficients are inverse quantized by the inverse quantization unit 160 and inversely transformed by the inverse transform unit 170. The inverse quantized and inverse transformed coefficients are added to the prediction block through an adder 175 and a reconstructed block is generated.

The reconstructed block passes through the filter unit 180, and the filter unit 180 may apply at least one of a deblocking filter and a sample adaptive offset (SAO) to a reconstructed block or a reconstructed picture. The filter unit 180 may also be referred to as an adaptive in-loop filter. The deblocking filter can remove block distortion occurring at the boundary between blocks. SAO may add an appropriate offset value to a pixel value to compensate for a coding error. The reconstructed block that has passed through the filter unit 180 is stored in the reference image buffer 190.

2 is a diagram illustrating an image decoding apparatus.

Referring to FIG. 2, the image decoding apparatus 200 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, a motion compensation unit 250, and a filter unit. 260) and a reference image buffer 270.

The image decoding apparatus 200 receives the bitstream output from the encoder, performs decoding in an intra mode or an inter mode, and outputs a reconstructed image, that is, a reconstructed image. In the case of intra mode, a prediction block is generated using an intra prediction mode, and in the case of an inter mode, a prediction block is generated using an inter prediction method. The image decoding apparatus 200 obtains a residual block from the received bitstream, generates a prediction block, and then adds the residual block and the prediction block to generate a reconstructed block, that is, a reconstructed block.

The entropy decoding unit 210 entropy-decodes the input bitstream according to a probability distribution and outputs a quantized coefficient. The quantized coefficients are inverse quantized by the inverse quantization unit 220 and inversely transformed by the inverse transform unit 230, and as a result of the inverse quantization/inverse transformation of the quantized coefficients, a residual block is generated.

In the case of the intra prediction mode, the intra prediction unit 240 (or the inter prediction unit) generates a prediction block by performing spatial prediction using pixel values of an already coded block around the current block.

In the case of the inter prediction mode, the motion compensation unit 250 generates a prediction block by performing motion compensation using a motion vector and a reference image stored in the reference image buffer 270.

The residual block and the prediction block are added through the adder 255, and the added block passes through the filter unit 260. The filter unit 260 may apply at least one of a deblocking filter and an SAO to a reconstructed block or a reconstructed picture. The filter unit 260 outputs a reconstructed image, that is, a reconstructed image. The reconstructed image may be stored in the reference image buffer 270 and used for inter prediction.

Methods for improving the prediction performance of an encoding/decoding apparatus include a method of increasing the accuracy of an interpolation image and a method of predicting a difference signal. Here, the difference signal is a signal representing the difference between the original image and the predicted image.

In the present invention, the'difference signal' may be replaced with'difference signal','residual block' or'difference block' depending on the context, and those of ordinary skill in the art may have an influence on the spirit and nature of the invention. You will be able to distinguish this within the scope not given.

In an embodiment of the present invention, for convenience of description, a coding unit (CU) is used as a term of a coding unit, but may be a unit that performs not only encoding but also decoding. Hereinafter, the video encoding method described in the embodiment of the present invention can be implemented in accordance with the function of each module, and such an encoder and a decoder are included in the scope of the present invention. That is, an image encoding/decoding method to be described later in an exemplary embodiment of the present invention may be performed by each component included in the image encoder and the image decoder. The meaning of the component may include not only a hardware meaning but also a software processing unit that can be performed through an algorithm.

In image encoding/decoding according to the exemplary embodiments of FIGS. 1 and 2, encoding is performed based on a basic unit of compression according to a standard for efficient encoding of an image. For example, the coding unit of H.264 is a macroblock, and in HEVC, a CU structure in which a coding unit of a single size is extended to various sizes may be defined.

In the present specification, a block or prediction coding unit means an encoding/decoding unit. In the encoding/decoding process, an image is divided into a predetermined size and encoded/decoded. In H.264, a block (prediction coding unit) may be referred to as a macroblock, and in HEVC, a block (prediction coding unit) may be referred to as CU, PU, TU, etc., and one block is a smaller-sized sub-block. It can also be divided.

The macroblock (MB) of the existing H.264, which is the basic unit of compression, has been replaced with a coding tree unit (CTU) in High-Efficiency Video Coding (HEVC). In this case, while the size of the MB has a size of 16x16, the size of the CTU can be selected from one of 16x16, 32x32, and 64x64. In addition, one CTU may be further divided into sub-coding units (CUs), and this CU is further divided into a prediction unit (PU) and a transform unit (TU).

Each CU may have a size of 64x64 to 8x8, a PU may have a size of 64x64 to 4x4, and a TU may have a size of 32x32 to 4x4. Therefore, HEVC supports block and frequency conversion of various sizes compared to H.264. In H.264, intra prediction (Intra-Prediction, intra prediction) has 9 prediction directions, whereas HEVC uses 35 prediction directions. In the case of inter-prediction (inter-prediction), the maximum block size is limited to 16x16 in the case of the existing H.264 inter prediction, whereas in HEVC, it has been extended to handle various block sizes from 4x4 to 64x64. .

A CU is a unit in which encoding is performed in a video encoder, and may be hierarchically divided with depth information based on a quad tree structure. CUs can have various sizes, such as 8Х8, 16Х16, 32Х32, 64Х64. In addition, the largest CU is referred to as LCU (Largest Coding Unit), and the smallest CU is referred to as SCU (Smallest Coding Unit). All CUs except the SCU allocate split_flag information to indicate whether the corresponding CU is a divided region or not according to the value. The encoder may adjust the size of the LCU in the encoding process according to various video signal characteristics. The CU may be divided into a prediction unit (PU) to be used for intra or inter prediction, and may be divided into a transform unit (TU) for transformation and quantization.

PU means a basic unit of performing prediction and/or motion compensation. The PU may be divided into a plurality of partitions, and each partition is referred to as a PU partition (Prediction Unit Partition). When the PU is divided into a plurality of partitions, the PU partition may be a basic unit for performing prediction and/or motion compensation. Hereinafter, in an embodiment of the present invention, PU may mean a PU partition.

High Efficiency Video Coding (HEVC) uses a Motion Vector Prediction method based on Advanced Motion Vector Prediction (AMVP).

In the motion vector prediction method based on the improved motion vector prediction, not only the motion vector (MV) of the reconstructed block located around the encoding/decoding target block, but also the same as the encoding/decoding target block within the reference picture. A motion vector of a block existing at a position or a corresponding position may be used. At this time. In the reference picture, a block existing at the same position or spatially corresponding to the encoding/decoding target block is a collocated block, and the motion vector of the equal position block is a collocated motion vector or temporal motion It is called a vector (Temporal Motion Vector). However, the collocated block is not only a block that always exists at the same position as the encoding/decoding target block of the reference picture, but may be a block that has a similar position to the encoding/decoding target block, that is, exists at a corresponding position. have.

In the motion information merge method, motion information is inferred not only from a neighboring reconstructed block but also from an equally positioned block, and used as motion information of an encoding/decoding target block. In this case, the motion information is inter prediction mode information indicating a reference picture index, motion vector, uni-direction or bi-direction required for inter prediction, and a reference picture list. ), information including at least one of prediction mode information on whether it is encoded in an intra prediction mode or an inter prediction mode.

The predicted motion vector in the encoding/decoding target block is not only a motion vector of a neighboring block spatially adjacent to the encoding/decoding target block, but also a motion vector of an equal position block that is a block temporally adjacent to the encoding/decoding target block. May be.

3 is a block diagram illustrating a scene change detection apparatus.

As illustrated in FIG. 3, the scene change detection apparatus 300 includes a candidate selection unit 310 and a scene determination unit 320. The scene change detection apparatus 300 may omit some of the various components exemplarily illustrated in FIG. 3 or may additionally include other components.

The scene change detection apparatus 300 calculates the ratio of intra prediction coding units by excluding the number of prediction coding units encoded in the skip mode from the total number of prediction coding units, so even if the frame sufficiently qualifies as a scene change candidate. Nevertheless, a problem in which skip mode prediction coding units due to low brightness cannot be detected is solved.

The candidate selector 310 calculates a ratio of a prediction coding unit encoded in an intra mode among all prediction coding units for each frame in the compression region. The candidate selector 310 selects a scene change candidate frame based on a ratio of prediction coding units encoded in the intra mode. The candidate selector 310 excludes the number of prediction coding units encoded in the skip mode from the total number of prediction coding units, and calculates a ratio of the prediction coding units encoded in the intra mode.

The scene determiner 320 detects a scene change using a complexity and an edge histogram based on an intra mode of a prediction coding unit encoded in an intra mode in the compressed region.

Since most of the coding modes of predictive coding units are coded in skip mode in an image frame having low brightness as a whole, if a scene change candidate is simply selected by the ratio of blocks coded in intra mode, overall low-light images such as night images or low-light images There is a problem with the accuracy of the candidates selected from the brightness image.

The candidate selection unit 310 detects a scene change candidate by improving a problem due to the number of blocks encoded in the skip mode. The candidate selection unit 310 selects a candidate frame according to a result of comparing the ratio of blocks encoded in the intra mode with a threshold value. When the candidate selection unit 310 primarily detects a scene change candidate frame, it calculates an intra-mode coding block ratio of the frame as shown in Equation (1).

The ratio (R ₁ ) of prediction coding units in the compression region encoded in the intra mode is from the number of prediction coding units (N _total ) belonging to the frame to the number of prediction coding units encoded in the skip mode (N _skip ). Is calculated. The ratio of the prediction coding units in the compression region encoded in the intra mode is the number of prediction coding units encoded in the inter mode (N _inter ) and the ratio of the prediction coding units encoded in the inter mode except for the prediction coding units encoded in the skip mode over the entire prediction coding units (N _inter ) The sum of the number of predicted coding units (N _intra ) is used as the denominator, and the number of prediction coding units (N _intra ) encoded in the intra mode is used as the numerator. The ratio of the prediction coding units in the compression region encoded in the intra mode is the number of prediction coding units (N _intra ) encoded in the intra mode and the number of prediction coding units encoded in the intra mode (N _intra ) for an image having a brightness value lower than the preset reference brightness value. It may be calculated by considering a relationship between the number of prediction coding units (N _skip ). A prediction coding unit (block) in H.264 may mean a macroblock, and a prediction coding unit (block) in HEVC may mean a PU.

In FIG. 4, a block divided by PU according to the HEVC prediction mode is illustrated. HEVC divides a block into Intra/Inter/Skip using a PU unit. In this embodiment, an intra mode coding block ratio may be calculated based on a PU unit for a video coded according to the HEVC standard. The ratio R ₂ of the prediction coding unit in the compression region encoded in the intra mode based on the PU unit is expressed as Equation 2.

In Equation 2, PU _total denotes the _total number of PUs in one frame, PU _skip denotes the number of PUs predicted in the skip mode in one frame, and PU _intra denotes the number of PUs predicted in the intra mode in one frame.

In the present embodiments, in the case of not only H.264 or HEVC, but also other compression standards, similarly to Equations 1 and 2, the coding ratio of the intra mode may be calculated by using the prediction coding unit in each standard as a measure. .

5 is an edge direction according to an intra prediction mode referenced by the scene change detection apparatus, and FIG. 6 is a diagram illustrating that the scene change detection apparatus constructs a complexity and edge histogram from one frame.

The scene determination unit 320 detects a scene change by using the edge histogram and the complexity extracted based on the intra mode of the block encoded in the intra mode in the compressed region.

The scene determiner 320 calculates the complexity according to the size of the intra mode of the prediction coding unit encoded in the intra mode. The scene determiner 320 determines that the smaller the prediction size of the intra prediction coding unit for each frame is, the higher the complexity, based on a tendency that a detailed portion of the image is encoded in a relatively small intra prediction coding unit.

The scene determiner 320 calculates an edge histogram according to the prediction direction of the intra mode of the prediction coding unit encoded in the intra mode. The scene determination unit 320 classifies an edge by referring to a prediction direction for each mode during intra prediction. The edge histogram contains non-edge bins instead of undirected edges.

The scene determination unit 320 uses a complexity and edge histogram difference (CEH), and the complexity and edge histogram difference is a histogram of the complexity and edge distribution of macroblocks belonging to each scene change candidate frame. It consists of bins. If the sum of the absolute values of the difference between the histogram bins between adjacent frames is greater than a preset threshold, the scene determination unit 320 determines that a scene change has occurred in the corresponding frame. A relational expression regarding the complexity and edge histogram difference is expressed as in Equation 3.

In Equation 3, CEHB _i (t) represents the value of the i-th histogram bin of the t-th scene change candidate frame. CEHB _i (t-1) represents the value of the I th histogram bin of the t-1 th scene change candidate frame. T _s is a preset threshold for detecting the final candidate for scene change. That is, the sum of the absolute values of the histogram bin differences corresponding to the same type between adjacent frames is taken and used for the final scene change detection.

The process of determining the optimal prediction mode method of HEVC is determined by rate-distortion optimization. This is determined as a mode in which the cost function is minimized by performing optimization of the cost function using the Lagrange Multiplier of Equation 4.

In Equation 4, J _mode is a cost function to be minimized. SSE is a Sum of Squared Error representing the error between blocks, ω _chroma is a weight for a color difference component, and λ _mode is a Lagrange coefficient. If SSE, ω _chroma , and λ _mode are specified, it is expressed as in Equation 5.

7 illustrates a process of determining a prediction mode for minimizing the cost function of Equation 4 according to each prediction mode. In the case of Skip mode, when the conditions of Early_Skip and CBF_Fast are satisfied for the first time, the optimal mode is immediately determined as Skip. In the case of Early_Skip, a condition is checked whether the coefficients of the residual blocks are all 0 and the motion vector difference is 0. If the Early_Skip condition is not satisfied, the CBF_Fast condition for checking whether the coefficient of the residual block is 0 is checked, and if not, Equation 4 for the Skip mode is calculated again. Thereafter, cost functions for other modes are calculated using Equation 4, the minimum cost function is found, and finally encoding is performed in the corresponding prediction mode.

In the case of a low-illuminance image at night, the block to be coded satisfies the Early_skip and CBF_Fast conditions, or the block resulting from Equation 2 in the Skip mode increases compared to the high-illuminance image in the daytime. It can be seen that excluding the Skip mode in the process of determining the transition can more accurately detect the scene transition at night.

Although components included in the scene change detection apparatus are shown separately in FIG. 3, a plurality of components may be combined with each other to be implemented as at least one module. Components are connected to a communication path connecting a software module or a hardware module inside the device and operate organically with each other. These components communicate using one or more communication buses or signal lines.

The scene change detection apparatus may be implemented in a logic circuit by hardware, firmware, software, or a combination thereof, or may be implemented using a general purpose or specific purpose computer. The device may be implemented using a hardwired device, a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), or the like. In addition, the device may be implemented as a System on Chip (SoC) including one or more processors and controllers.

The scene change detection apparatus may be mounted in software, hardware, or a combination thereof on a computing device or server provided with a hardware element. Computing devices or servers include all or part of a communication device such as a communication modem for performing communication with various devices or wired/wireless communication networks, a memory storing data for executing a program, and a microprocessor for calculating and commanding a program. It can mean various devices including.

8 is a flowchart illustrating a method for detecting a scene change according to another embodiment of the present invention. The scene change detection method may be performed by the computing device, and a detailed description of an operation performed by the scene change detection apparatus and a duplicate description will be omitted.

In step S810, the computing device calculates a ratio of the prediction coding units encoded in the intra mode from the frames of the image captured at night in the compression region, and selects a scene change candidate using the ratio of the prediction coding units encoded in the intra mode. do. The ratio of the prediction coding units encoded in the intra mode is calculated by excluding the number of prediction coding units encoded in the skip mode from all the prediction coding units belonging to the frame. The ratio of the prediction coding units encoded in the intra mode is the sum of the number of prediction coding units encoded in the inter mode and the number of prediction coding units encoded in the intra mode excluding the prediction coding units encoded in the skip mode from all the prediction coding units. Is the denominator, and the number of prediction coding units encoded in the intra mode is taken as the numerator. The ratio of the prediction coding units encoded in the intra mode is the relationship between the number of prediction coding units encoded in the intra mode and the number of prediction coding units encoded in the skip mode for an image having a brightness value lower than a preset reference brightness value. Can be calculated by taking into account. In the step S810 of selecting a scene change candidate, a candidate frame is selected according to a result of comparing the ratio of the prediction coding units encoded in the intra mode with a threshold value.

In step S820, the computing device detects a scene change by using the edge histogram and the complexity extracted based on the intra mode of the prediction coding unit encoded in the intra mode in the compression region. In the step of detecting a scene change (S820), the complexity is calculated according to the size of the intra mode of the prediction coding unit encoded in the intra mode, and the edge histogram is calculated according to the prediction direction of the intra mode of the prediction coding unit encoded in the intra mode. do. The edge histogram may include non-edge bins.

9 is a diagram illustrating a macroblock mode distribution of a scene among night black box captured images processed by a scene change detection apparatus and method according to embodiments of the present invention.

The frame is a scene in which the bus continues to appear in the previous frames and is passing the car after, and corresponds to the scene change. In the image, a macroblock indicated in red is an intra mode macroblock encoded in an intra-mode mode, blue is an inter mode macroblock encoded in an inter-screen mode, and black is a skip mode macroblock. The percentage of each mode is 36.85%, 26.12%, and 37.03%, respectively. Since the ratio of intra mode macroblocks calculated including the skip mode is 36.85% and is smaller than the threshold (eg, 50%), it is not selected as a candidate. However, if the first scene change candidate detection equation is applied according to the embodiments, the ratio of all macroblocks excluding skip mode macroblocks to intra mode macroblocks (36.85/(36.85+26.12)*100=58.52%) is critical Since it becomes sufficiently larger than the value (eg, 50%), the frame may be detected as a scene change candidate.

In the present embodiments, a ratio of an intra-screen coded block is calculated for each frame, a scene change candidate is selected based on the ratio of the intra-screen coded blocks, and the number of blocks coded in the skip mode is determined as the total block. By excluding from the number of and detecting scene changes using the complexity and edge histograms based on intra prediction in the compressed region, it is not very dependent on the syntax of the H.264 or HEVC standard, which is a high-definition video compression standard, and is relatively syntax independent. It can be easily and flexibly implemented within the existing standard, and finally a frame close to the actual scene change can be detected through the scene change evaluation criteria robust to the brightness of the image without complete decoding of the image in the compressed region.

In FIG. 8, each process is described as sequentially executing, but this is only illustrative, and those skilled in the art may change the order shown in FIG. 8 within the scope not departing from the essential characteristics of the embodiment of the present invention. Or, by executing one or more processes in parallel, or adding other processes, various modifications and variations may be applied.

The operations according to the embodiments may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. Computer-readable medium refers to any medium that has participated in providing instructions to a processor for execution. The computer-readable medium may include program instructions, data files, data structures, or a combination thereof. For example, there may be a magnetic medium, an optical recording medium, a memory, and the like. Computer programs may be distributed over networked computer systems to store and execute computer-readable codes in a distributed manner. Functional programs, codes, and code segments for implementing the present embodiment may be easily inferred by programmers in the technical field to which the present embodiment belongs.

These embodiments are for explaining the technical idea of the present embodiment, and the scope of the technical idea of the present embodiment is not limited by these embodiments. The scope of protection of this embodiment should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present embodiment.

300: scene change detection device
310: Candidate elections
320: scene determination unit

Claims (13)

In the method for detecting a scene change of a night image by a computing device,
Calculating a ratio of prediction coding units in a compressed region encoded in intra mode from a frame of an image captured at night, and selecting a scene change candidate using the ratio of prediction coding units in the compressed region encoded in the intra mode step; And
And detecting a scene change using an edge histogram and a complexity extracted based on an intra mode of a prediction coding unit in the compression region encoded in the intra mode,
The ratio of prediction coding units in the compression region encoded in the intra mode is,
For an image having a brightness value lower than a preset reference brightness value, the image is calculated by considering a relationship between the number of prediction coding units encoded in the intra mode and the number of prediction coding units encoded in the skip mode. Scene change detection method. The method of claim 1,
The prediction coding unit in the compression region means a macroblock of H.264 or a prediction unit of HEVC,
A scene change of a night image, characterized in that the ratio of prediction coding units in the compression region encoded in the intra mode is calculated by excluding the number of prediction coding units encoded in the skip mode from all prediction coding units belonging to the frame Detection method. The method of claim 1,
The ratio of prediction coding units in the compression region encoded in the intra mode is,
In all prediction coding units, the sum of the number of prediction coding units encoded in the inter mode and the number of prediction coding units encoded in the intra mode excluding the prediction coding units encoded in the skip mode is used as the denominator, and the intra mode A method for detecting a scene change of a night image, characterized in that the number of predictive coding units encoded as a numerator is used. delete The method of claim 1,
The step of selecting the scene change candidate,
And selecting a candidate frame based on a result of comparing a ratio of the prediction coding units encoded in the intra mode with a threshold value. The method of claim 1,
The step of detecting the scene change,
The complexity is calculated according to the size of the intra mode of the prediction coding unit in the compression region encoded in the intra mode, and the edge histogram according to the prediction direction of the intra mode of the prediction coding unit in the compression region encoded in the intra mode And the edge histogram includes a non-edge bin. Calculating a ratio of prediction coding units in a compressed region encoded in intra mode from a frame of an image captured at night, and selecting a scene change candidate using the ratio of prediction coding units in the compressed region encoded in the intra mode Candidate elections; And
A scene determination unit for detecting a scene change using a complexity and an edge histogram extracted based on an intra mode of a prediction coding unit in the compression region encoded in the intra mode,
The ratio of prediction coding units in the compression region encoded in the intra mode is,
For an image having a brightness value lower than a preset reference brightness value, the image is calculated by considering a relationship between the number of prediction coding units encoded in the intra mode and the number of prediction coding units encoded in the skip mode. Scene change detection device. The method of claim 7,
The prediction coding unit in the compression region means a macroblock of H.264 or a prediction unit of HEVC,
A scene change of a night image, characterized in that the ratio of prediction coding units in the compression region encoded in the intra mode is calculated by excluding the number of prediction coding units encoded in the skip mode from all prediction coding units belonging to the frame Detection device. The method of claim 7,
The ratio of prediction coding units in the compression region encoded in the intra mode is,
In all prediction coding units, the sum of the number of prediction coding units encoded in the inter mode and the number of prediction coding units encoded in the intra mode excluding the prediction coding units encoded in the skip mode is used as the denominator, and the intra mode A scene change detection apparatus for a night image, characterized in that the number of predictive coding units encoded as numerators is used as a numerator. delete The method of claim 7,
The candidate election unit,
And selecting a candidate frame according to a result of comparing a ratio of a prediction coding unit in a compression region encoded in the intra mode with a threshold value. The method of claim 7,
The scene determination unit,
The complexity is calculated according to the size of the intra mode of the prediction coding unit in the compression region encoded in the intra mode, and the edge histogram according to the prediction direction of the intra mode of the prediction coding unit in the compression region encoded in the intra mode And the edge histogram includes a non-edge bin. A computer program for detecting a scene change of a night image by being recorded in a non-transitory computer-readable medium including computer program instructions executable by a processor, the computer program instructions being stored in at least one processor of a computing device. If implemented by,
Calculating a ratio of prediction coding units in a compressed region encoded in intra mode from a frame of an image captured at night, and selecting a scene change candidate using the ratio of prediction coding units in the compressed region encoded in the intra mode step; And
Performing operations including detecting a scene change using an edge histogram and a complexity extracted based on an intra mode of a prediction coding unit in a compression region encoded in the intra mode,
The ratio of prediction coding units in the compression region encoded in the intra mode is,
A computer program, characterized in that, for an image having a brightness value lower than a preset reference brightness value, it is calculated by considering a relationship between the number of prediction coding units encoded in the intra mode and the number of prediction coding units encoded in the skip mode.

Images