KR100404328B1 - Method for detecting motion vector of dynamic image - Google Patents

Method for detecting motion vector of dynamic image Download PDF

Info

Publication number
KR100404328B1
KR100404328B1 KR1019960018275A KR19960018275A KR100404328B1 KR 100404328 B1 KR100404328 B1 KR 100404328B1 KR 1019960018275 A KR1019960018275 A KR 1019960018275A KR 19960018275 A KR19960018275 A KR 19960018275A KR 100404328 B1 KR100404328 B1 KR 100404328B1
Authority
KR
South Korea
Prior art keywords
motion vector
difference value
value
image
current
Prior art date
Application number
KR1019960018275A
Other languages
Korean (ko)
Other versions
KR970076341A (en
Inventor
김정현
Original Assignee
주식회사 하이닉스반도체
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 주식회사 하이닉스반도체 filed Critical 주식회사 하이닉스반도체
Priority to KR1019960018275A priority Critical patent/KR100404328B1/en
Publication of KR970076341A publication Critical patent/KR970076341A/en
Application granted granted Critical
Publication of KR100404328B1 publication Critical patent/KR100404328B1/en

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • H04N19/513—Processing of motion vectors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/223—Analysis of motion using block-matching
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/136—Incoming video signal characteristics or properties
    • H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
    • H04N19/139—Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

PURPOSE: A method for detecting a motion vector of a dynamic image is provided to reduce time consumption required for full search, thereby promptly detecting a motion vector. CONSTITUTION: A reference image which is a previous picture, and a current input image are inputted to a motion estimator(200), and search windows are formed in the reference image for finding a motion vector by macro blocks. An absolute error calculator(210) receives macro block output from the motion estimator for adding up all least absolute error values to output a mean value. The absolute error calculator transmits the mean value to a comparator(220). The comparator stops detecting the motion vector if a condition related to difference values of macro blocks are satisfied. The condition is either a value obtained by subtracting a difference value of the immediately previous macro block from a difference value of a current macro block is smaller than one tenth of the difference value of the current macro block, or the difference value of the current macro block is smaller than an error limit value.

Description

동화상의 움직임 벡터 검출방법Motion vector detection method of motion picture

본 발명은 동화상의 움직임 벡터 검출방법에 관한 것으로, 특히 풀서치(Full search)의 시간소모를 줄이기 위하여 주어진 조건을 만족하면 더 이상 검출하지 않도록 함으로써, 움직임 벡터를 빠르게 검출하는 데에 적당하도록 한 동화상의 움직임 벡터 검출방법에 관한 것이다.The present invention relates to a motion vector detection method for a moving picture, and more particularly, to a motion vector detection method for detecting a motion vector in a moving picture suitable for quickly detecting a motion vector, To a motion vector detection method.

제1도는 일반적인 프레임간 부호화기의 블럭도로서, 가산기(100)에서는 현재입력되는 영상과 움직임 보상부(110)로부터 출력되는 움직임이 보상된 영상을 입력받아 이의 차영상을 출력하게 되며, 이 차영상은 이산여현변환부(101)를 통해 이산여현변환된 후 양자화부(102)를 통해 양자화된다.FIG. 1 is a block diagram of a general inter-frame encoder. In the adder 100, a current input image and a motion compensated image output from the motion compensating unit 110 are received and a difference image is output. Are subjected to DCT through a DCT unit 101 and quantized through a quantization unit 102. [

부호화부(103)에서는 상기 양자화부(102)의 출력을 입력받아 가변길이 및 고정길이 부호화를 행한 다음 버퍼(104)에 저장된다.The encoding unit 103 receives the output of the quantization unit 102, performs variable length and fixed length encoding, and then stores the encoded data in the buffer 104.

한편, 상기 양자화부(102)의 출력은 역양자화부(105) 및 역이산여환변환부(106)를 통하여 차영상이 다시 복원되고, 이 차영상은 가산기(107)에 입력되어 움직임 보상부(110)로부터 출력되는 영상신호와 합쳐져 프레임 메모리(108)에 저장된다.The output of the quantization unit 102 is reconstructed again through an inverse quantization unit 105 and an inverse discrete cosine transform unit 106. The difference image is input to an adder 107 to be input to a motion compensation unit 110, and stored in the frame memory 108. The frame memory 108 stores the image signals.

이 과정은 프레임간 복호화기에서도 그대도 적용되며, 상기에서 기술한 방법으로 영상이 복원된다.This process is also applied to the inter-frame decoder, and the image is reconstructed by the method described above.

움직이 보상부(110)에서는 이렇게 복원되어 프레임 메모리(108)에 저장된 영상을 기준영상으로 하고, 움직임 추정부(109)의 움직임 벡터(Motion vector)를 이용하여 움직임이 보상된 영상을 구성하게 된다.The motion compensation unit 110 reconstructs the reference image using the image stored in the frame memory 108 and constructs the motion compensated image using the motion vector of the motion estimation unit 109 .

여기서, 상기 움직임 보상부(110)에서 사용하는 움지임 벡터는 상기 움직임 추정부(109)에서 프레임 메모리(108)의 저장된 복원영상을 기준영상으로 하여 현재 입력되어지는 영상의 움직임을 예측영상으로 삼아 구해지게 된다.Here, the motion vector used in the motion compensation unit 110 is a motion vector estimated by the motion estimation unit 109 using the stored reconstructed image of the frame memory 108 as a reference image, .

그리고, 제2도는 상기 움직임 추정부(109)에서의 움직임 벡터를 검출하는 방법을 예시한 도로, 현재화면의 움직임을 찾고자 하는 예측영상의 매크로(Macro) 블럭(MB)과 전화면 즉, 프레임 메모리(108)에 저장된 기준영상의 매크로블럭(16, 16픽셀)을 비교함으로써, 가장 닮은 블럭의 상대적 위치를 찾는 것이다.FIG. 2 is a diagram illustrating a method of detecting a motion vector in the motion estimation unit 109, a macro block (MB) of a predictive image to be searched for a motion of the current screen, (16, 16 pixels) of the reference image stored in the reference block 108 to find the relative position of the most similar block.

이는 기준영상을 가지고 탐색영역상에서 현재 움직임을 찾고자하는 블럭을 화소 단위로 수평 및 수직 방향으로 이동시켜 가면서 기준영상과의 차이를 구해 가장 차이가 적을때의 상대적인 좌표(X,Y)를 움직임 벡터로 구하는 것이다.In this case, a block to find the current motion on the search area with reference images is shifted in the horizontal and vertical directions on a pixel-by-pixel basis, and the difference between the reference image and the reference image is obtained. The relative coordinates (X, Y) .

이때, 현재화면의 찾고자하는 매크로블럭에 해당하는 좌표에 대해 전화면 기준영상의 매크로블럭의 좌표를 중심으로 탐색창(Search window)을 결정하고, 전화면 기준영상의 탐색창을 매크로블럭으로 나눈 뒤 현재화면의 매크로블럭과 모두 비교한다.At this time, a search window is determined on the basis of the coordinates of the macroblock of the full screen reference image with respect to the coordinates corresponding to the macroblock to be searched on the current screen, the search window of the full screen reference image is divided into macroblocks Compare with the current macroblock.

그러나, 이와같은 종래의 방법은 풀서치(Full search)의 방법을 이용하여 움직임 벡터를 검출해야 하므로 시간이 많이 소모되는 문제를 갖는다.However, such a conventional method has a problem that a time is consumed because a motion vector must be detected using a full search method.

본 발명은 이와같은 종래의 문제를 해결하기 위하여 창안된 것으로, 풀서치의 시간소모를 줄이기 위해 매크로블럭의 차이값에 대한 조건을 주어, 이 조건을 만족하게 되면 더 이상 움직임 벡터를 찾지 않도록 하는 동화상의 움직임 벡터 검출방법을 제공함에 그 목적이 있다.SUMMARY OF THE INVENTION The present invention has been made in order to solve the conventional problems as described above. In order to reduce the time consumption of the full search, And a method of detecting a motion vector of the motion vector.

상기와 같은 목적을 달성하기 위한 본 발명 동화상의 움직임 벡터 검출방법은 움직임 추정부(200)에 전화면인 기준영상과 현재 입력영상의 데이타가 입력되고, 상기 전화면인 기준영상에서 탐색창을 구성하여 매크로블럭 단위로 찾는 제1과정과, 절대오차 계산부(210)에서 상기 제1과정의 매크로블럭 출력을 입력받아 각 탐색창에서 찾아낸 절대오차 값 중 가장 적은 값들을 모두 더해 평균을 내는 제2과정과, 상기 제2과정의 출력값들을 스위칭하여 비교부(220)로 보내고 이비교부(220)에서의 조건을 만족하면 움직임 벡터의 검출을 멈추는 제3과정으로 이루어지는 것으로, 이와같은 본 발명 동화상의 움직임 벡터의 검출방법에 대해 제3도를 참조하여 상세히 설명한다.According to another aspect of the present invention, there is provided a method for detecting motion vectors of moving pictures, comprising: inputting data of a reference image and a current input image, which are full-screened, to a motion estimation unit, A macroblock output of the first process is input to the absolute error calculator 210, and the absolute value of the absolute value of the macroblock output of the second process And a third step of switching output values of the second process to the comparison unit 220 and stopping the detection of a motion vector if the condition of the comparison unit 220 is satisfied. A method of detecting a motion vector will be described in detail with reference to FIG.

본 발명은, 종래의 동화상의 움지임 벡터 검출방법과 비교해 볼때, 풀서치의 시간소모를 줄이기 위해 하기의 조건을 만족하면 움직임 벡터를 더 이상 찾지 않는 데에 그 차이가 있다.The present invention differs from the conventional motion vector detection method in that motion vectors are not found any more if the following conditions are satisfied in order to reduce the time consumption of the full search.

이때, (n)은 매크로블럭(MB)의 차이값이고, E(2)는 현재위치에서의 차이값이며, E(1)은 현재 바로 이전에서의 차이값이다.Here, (n) is the difference value of the MB (MB), E (2) is the difference value at the current position, and E (1) is the difference value immediately before the current MB.

먼저, 움직임 추정부(200)에 현재 입력영상과 과거의 기준영상 데이타가 입력되면, 과거의 기준영상에서 탐색창을 구성하여 매크로블럭 단위로 찾은 뒤에 절대오차 계산부(210)에 입력된다.First, when a current input image and a past reference image data are inputted to the motion estimation unit 200, a search window is formed in a past reference image, and the search window is searched for in units of macroblocks and then input to an absolute error calculation unit 210.

절대오차 계산부(210)에서는 각 탐색창에서 찾아낸 절대오차 값 중 가장 작은 값들을 모두 더해 평균을 낸 뒤 스위칭시켜 비교부(220)로 보낸다.The absolute error calculator 210 adds all the smallest absolute error values found in each search window, averages them, and then sends the average values to the comparator 220.

비교부(220)에서는 상기 언급한 조건들에 만족되면 즉, 현재 매크로블럭(MB)의 차이값 E(2)에서 바로 이전 매크로블럭(MB)의 차이값 E(1)을 뺀 값이 상기 차이값 E(2)의 1/10보다 작거나, 상기 현재 매크로블럭(MB)의 차이값 E(2)이 오차한계값(Eth)보다 작은 경우에 움직임 벡터를 찾는 일을 더 이상 하지 않게 된다.The comparator 220 compares the value obtained by subtracting the difference value E (1) of the immediately preceding macroblock (MB) from the difference value E (2) of the current macroblock (MB) It is no longer necessary to find a motion vector when the difference value E (2) of the current macroblock (MB) is smaller than 1/10 of the value E (2) or smaller than the error threshold value Eth.

상술한 바와 같이 본 발명 동화상의 움직임 벡터 검출방법에 의하면, 매크로블럭의 차이값에 대한 조건을 주어서 이 조건을 만족하면 움직임 벡터 검출을 더 이상 하지 않도록 하여 풀서치의 시간소모를 줄일 수 있게 함으로써, 움직임 벡터를 빠르게 검출할 수 있는 효과가 있다.As described above, according to the motion vector detection method of the present invention, when the condition for the difference value of the macroblock is satisfied and the condition is satisfied, the motion vector detection is no longer performed, and the time consumption of the full search can be reduced, There is an effect that a motion vector can be detected quickly.

제1도는 일반적인 프레임간 부호화장치를 나타낸 블럭도.FIG. 1 is a block diagram showing a typical inter-frame encoding apparatus. FIG.

제2도는 움직임 벡터의 검출방법을 나타낸 것으로,FIG. 2 shows a method of detecting a motion vector,

(가)는 전화면의 매크로블럭을 나타낸 도이고,(A) is a diagram showing a macro block of the entire screen,

(나)는 현재화면의 매크로블럭을 나타낸 도.(B) is a diagram showing a macroblock of the current screen.

제3도는 본 발명 동화상의 움직임 벡터 검출장치의 블럭도.FIG. 3 is a block diagram of a motion vector detecting apparatus according to the present invention; FIG.

★★★ 도면의 주요 부분에 대한 부호의 설명 ★★★Description of Reference Numerals for Main Parts of the Drawing

200 : 움직임 추정부 210 : 절대오차 계산부200: motion estimation unit 210: absolute error calculation unit

220 : 비교부220:

Claims (1)

움직임 추정부에 이전화면인 기준영상과 현재 입력영상의 데이타가 입력되고, 상기 이전화면인 기준영상에서 탐색창을 구성하여 매크로블럭 단위로 찾는 제1과장과, 절대오차 계산부에서 상기 제1과정의 매크로블럭 출력을 입력받아 각 탐색창에서 찾아낸 절대오차 값 중 가장 적은 값들을 모두 더해 평균을 내는 제2과정과, 상기 제2과정의 출력값들을 스위칭하여 비교부로 보내고 이 비교부에서 현재 매크로블럭의 차이값에서 바로 이전 매크로블럭의 차이값을 뺀 값이 상기 현재 매크로블럭이 차이값의 1/10보다 작거나, 상기 현재 매크로블럭의 차이값이 오차 한계값 보다 작은 경우에 움직임 벡터의 검출을 멈추는 제3과정으로 이루어진 것을 특징으로 하는 동화상의 움직임 벡터 검출방법.A first exaggeration unit for inputting data of a reference image and a current input image, which are previous images, into a motion estimation unit, and a search window in a reference image, which is a previous image, in units of macroblocks; A second step of receiving the output of the macroblock of the first macroblock and adding the smallest absolute value among the absolute error values found in each search window and averaging the output, and outputting the output values of the second process to the comparator, If the value obtained by subtracting the difference value of the immediately preceding macroblock from the difference value is less than 1/10 of the difference value of the current macroblock or the difference value of the current macroblock is smaller than the error threshold value, And a third step of detecting the motion vector of the moving picture.
KR1019960018275A 1996-05-28 1996-05-28 Method for detecting motion vector of dynamic image KR100404328B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1019960018275A KR100404328B1 (en) 1996-05-28 1996-05-28 Method for detecting motion vector of dynamic image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019960018275A KR100404328B1 (en) 1996-05-28 1996-05-28 Method for detecting motion vector of dynamic image

Publications (2)

Publication Number Publication Date
KR970076341A KR970076341A (en) 1997-12-12
KR100404328B1 true KR100404328B1 (en) 2004-02-05

Family

ID=37422592

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019960018275A KR100404328B1 (en) 1996-05-28 1996-05-28 Method for detecting motion vector of dynamic image

Country Status (1)

Country Link
KR (1) KR100404328B1 (en)

Also Published As

Publication number Publication date
KR970076341A (en) 1997-12-12

Similar Documents

Publication Publication Date Title
KR100275694B1 (en) Hierarchical search block matching method by using multiple motion vector candidates
US6256343B1 (en) Method and apparatus for image coding
US6360017B1 (en) Perceptual-based spatio-temporal segmentation for motion estimation
KR100242406B1 (en) Method for motion estimation using trajectory in a digital video encoder
JPH0870460A (en) Movement compensation type coding method adapted to magnitude of movement,and its device
KR20010083717A (en) Motion estimation method and appratus
KR100460950B1 (en) Transcoder and transcoding method
KR20040089163A (en) Method and Apparatus for Determining Search Range for Adaptive Motion Vector for Use in Video Encoder
US8155459B2 (en) Video processing device with low memory bandwidth requirements
JPH06311502A (en) Motion picture transmission equipment
KR100262500B1 (en) Adaptive block effect reduction decoder
KR100286818B1 (en) Fast motion estimating method for real-time video coding
KR100404328B1 (en) Method for detecting motion vector of dynamic image
JP2009284058A (en) Moving image encoding device
KR19990058255A (en) Video compression method using 1/4 pixel motion vector.
KR100295327B1 (en) Motion Estimation System Using Adaptive Threshold
JP2003061099A (en) Motion detection method in encoder
KR100382540B1 (en) Motion estimation method
JPH06296276A (en) Pre-processor for motion compensation prediction encoding device
KR100290618B1 (en) Method for estimating motion of block unit in image compression coding
KR100323660B1 (en) Scene adaptive motion estimator of hdtv
KR100617177B1 (en) Motion estimation method
KR100262962B1 (en) Motion prediction method and apparatus
KR20000037945A (en) Apparatus for efficiently tracing motion considering contour of image
Duanmu et al. A continuous tracking algorithm for long-term memory motion estimation [video coding]

Legal Events

Date Code Title Description
A201 Request for examination
N231 Notification of change of applicant
E902 Notification of reason for refusal
E701 Decision to grant or registration of patent right
GRNT Written decision to grant
LAPS Lapse due to unpaid annual fee