Abstract:
이진화 파라메터를 업데이트하여 변환 계수 레벨 정보를 복호화하는 방법 및 장치가 개시된다. 비디오 복호화 방법은 비트스트림에 대해 컨텍스트 모델에 기초한 산술 복호화를 수행함으로써 현재 변환 계수 레벨 정보에 대응하는 비트열을 획득하고, 이전 이진화 파라메터에 기초하여 획득된 소정 임계값과 이전 변환 계수의 크기를 비교한 결과에 기초하여, 상기 이전 이진화 파라메터를 갱신하거나 유지함으로써, 현재 이진화 파라메터를 결정하며, 상기 현재 이진화 파라메터를 이용하여 상기 비트열을 역이진화함으로써 상기 현재 변환 계수 레벨 정보를 획득하고, 상기 현재 변환 계수 레벨 정보를 이용하여 현재 변환 계수의 크기를 생성한다. 소정 임계값보다 상기 이전 변환 계수의 크기가 큰 경우, 상기 이전 이진화 파라메터에 1을 더한 값이 상기 현재 이진화 파라메터로 결정된다.
Abstract:
Disclosed in the present invention are a method of determining a reference picture for inter-prediction and an inter-prediction method according to the same. More specifically, the inter-prediction method is characterized in that a reference picture list which is used by a current prediction unit is determined when a current slice is a B slice; inter-prediction index information of the current prediction unit which indicates one of an L0 list and an L1 list is outputted when the size of the current prediction unit is 4x8 or 8x4; and inter-prediction index information which indicates a reference picture list of the L0 list, the L1 list, or both lists is outputted when the size of the current prediction unit is neither 4x8 nor 8x4.
Abstract:
The present invention suggests a method for determining a reference image for inter prediction and an inter prediction method thereof. The present invention provides an inter prediction method capable of determining a reference picture list using a current prediction unit when a current slice is a B slice, outputting inter prediction index information of the current prediction unit indicating one reference picture list among a L0 list and L1 list when the size of the current prediction unit is 4X8 or 8X4, and outputting one reference picture list among the L0 list, the L1 list, and a pair list when the size of the current unit is not 4X8 or 8X4. [Reference numerals] (12) Reference picture list determining part; (14) Reference index determining part
Abstract:
Disclosed are methods for generating and playing a 3D image data stream. A first portion image which comprises half of the data of a 3D image, and a second portion image which comprises the other half of the data of the 3D image are coded, wherein the 3D image comprises a first view image having full resolution and a second view image having full resolution. The coded image data is included in one stream or two streams and transmitted, on the basis of a first stream generation technique or a second stream generation technique. Information on the stream generation technique, information on a partial image to which image data included in a current stream is included, and view information are included in the information stream. [Reference numerals] (11) 3D image multiplexing unit; (13) First image coding unit; (14) Second image coding unit; (16) Image data stream generating unit; (17) Information stream generating unit; (AA) First view image; (BB) First portion image; (CC) Second view image; (DD) Second portion image
Abstract:
PURPOSE: A parameter updating method for encoding and decoding the entropy of a conversion factor level, an entropy encoding apparatus of the conversion factor level using thereof, and an entropy decoding apparatus thereof are provided to prevent the parameter, which is used when encoding and decoding the entropy of the conversion factor level, from radically changing due to a large size of conversion factor. CONSTITUTION: A parameter determining unit (2610) determines the renewal of the previous parameter. By renewing or maintaining the previous parameter based on the determination result, the parameter determining unit obtains the parameter, which is used for the inverse binary of the current conversion factor level syntax element showing the size of the current conversion factor. A syntax element restoring unit (2620) obtains the size of current conversion factor by enabling the inverse binary of the current conversion factor level syntax element by using the obtained parameter. [Reference numerals] (2610) Parameter determining unit; (2620) Syntax element restoring unit
Abstract:
PURPOSE: A video encoding method for parallel processing, an apparatus for the same, a video decoding method thereof and an apparatus for the video decoding method are provided to express whether multiple parallel processing types for parallel processing of video data are used or not through a unified syntax. CONSTITUTION: A parallel processing information output unit (1520) encodes parallel processing syntax information which represents a parallel processing type applied to a second data unit in a first data unit header. The first data unit header includes encoding information of a first data unit. The parallel processing information output unit encodes a parallel processing flag which represents whether the parallel processing type indicated by a parallel processing syntax is applied to the second data unit in a second data unit header. The second data unit header includes encoding information of the second data unit. [Reference numerals] (1510) Parallel processing judgment unit; (1520) Parallel processing information output unit
Abstract:
PURPOSE: A method for determining a quantization parameter based on a conversion block size and a device thereof are provided to reduce a quantization error by allocating a small quantization parameter according to an increase in the size of a conversion unit. CONSTITUTION: A conversion unit determining unit determines one or more conversion units included in a current encoding unit(21). A quantization parameter determining unit determines a basic quantization parameter for the current encoding unit(23). The quantization parameter determining unit reduces a quantization parameter for a conversion unit greater than a predetermined size(25). The quantization parameter determining unit increases a quantization parameter for a conversion unit smaller than the predetermined size(27). [Reference numerals] (21) Determine conversion units included in an encoding unit; (23) Determine the basic quantization parameter for the encoding unit; (25) Reduce a quantization parameter, which is larger than a predetermined size, compared to the basic quantization parameter; (27) Increase a quantization parameter, which is smaller than a predetermined size, compared to the basic quantization parameter; (AA) Start; (BB) End
Abstract:
A method for accessing a memory in a video processing device is provided to enable the minimum number of memory accesses during motion compensation operation to have more efficient operation capability, thereby easily performing an H.264 decoder in real time. A method for accessing a memory comprises the following steps of: portioning a macro block and load corresponding 8x8 block pixels and residual pixel data for inter prediction from a frame buffer, and storing and using the loaded 8x8 block pixels and residual pixel data in performing the inter prediction of an 8x8 block; and reusing the pixel data stored in a start memory in the inter prediction of the 8x8 block and sub blocks.
Abstract:
An apparatus and a method for discriminating a filter condition region in a deblocking filter are provided to reduce the number of cycles generated in calculating boundary strength for a macro block in the deblocking filter, thereby processing an image in real time. A memory unit(602) stores a table having a filter condition region value for boundary strength calculation corresponding to each of a type value of a current macro block and a type value of a neighboring macro block and stores a boundary strength calculation method by filter condition region. A deblocking filter(605) performs corresponding boundary strength calculation according to a filter condition region for an input macro block. A control unit(600) checks a filter condition region value corresponding to the type value of the current macro block and the type value of the neighboring macro block by using the table in the calculation of the boundary strength in the current macro block inputted to the deblocking filter, and controls the filtering of the deblocking filter according to the checked filter condition region value.