Method and device for determining similarity between sequences

Patent metadata

Field	Value
Publication number	US-2016267353-A1
Application number	US-201615016492-A
Country	US
Kind code	A1
Filing date	Feb 5, 2016
Priority date	Mar 10, 2015
Publication date	Sep 15, 2016
Grant date	—

Abstract

A method and a device for determining similarity between sequences are provided. The method includes obtaining a first sequence of frames and a second sequence of frames, determining a descriptor of a first frame of the first sequence based on luma information of blocks in the first frame, the descriptor of the first frame including luma difference information that is determined based on the luma information of the blocks in the first frame, determining a descriptor of a second frame of the second sequence based on luma information of blocks in the second frame, the descriptor of the second frame including luma difference information that is determined based on the luma information of the blocks in the second frame, and determining a similarity between the first sequence and the second sequence based on the descriptor of the first frame and the descriptor of the second frame.

First claim

What is claimed is:

1 . A method of determining similarity between sequences, the method comprising:
obtaining a first sequence of frames and a second sequence of frames; determining a descriptor of a first frame of the first sequence based on luma information of blocks in the first frame, the descriptor of the first frame comprising luma difference information that is determined based on the luma information of the blocks in the first frame; determining a descriptor of a second frame of the second sequence based on luma information of blocks in the second frame, the descriptor of the second frame comprising luma difference information that is determined based on the luma information of the blocks in the second frame; and determining a similarity between the first sequence and the second sequence based on the descriptor of the first frame and the descriptor of the second frame.

2 . The method of claim 1 , wherein the determining the similarity comprises:
matching the first frame with the second frame based on the descriptor of the first frame and the descriptor of the second frame; and determining a section in the second sequence that matches a section in the first sequence as a section similar to the section in the first sequence, based on the matched first frame and second frame.

3 . The method of claim 2 , wherein the matching comprises:
determining whether a Hamming distance between the descriptor of the first frame and the descriptor of the second frame is shortest; and matching the first frame with the second frame in response to the determining that the Hamming distance between the descriptor of the first frame and the descriptor of the second frame is shortest.

4 . The method of claim 2 , wherein the determining the section in the second sequence comprises:
obtaining a matching vector representing a matching relation between the matched first frame and second frame; determining a representative matching vector based on the obtained matching vector; and determining the section in the second sequence that matches the section in the first sequence as the section similar to the section in the first sequence, based on the representative matching vector.

5 . The method of claim 4 , further comprising:
determining whether the matching vector of the first frame is not included in a range based on the representative matching vector; and re-matching the first frame with a frame in the second sequence within the range based on the representative matching vector, in response to the determining that the matching vector of the first frame is not included in the range based on the representative matching vector.

6 . The method of claim 1 , further comprising:
extracting one or more first frames from the frames of the first sequence based on a frame rate, the first frame being one of the one or more first frames; and extracting one or more second frames from the frames of the second sequence based on the frame rate, the second frame being one of the one or more second frames.

7 . The method of claim 1 , wherein the determining the descriptor of the first frame or the second frame comprises:
determining whether a number of pixels having a same pixel value at a same location in the frames of the first sequence or the second sequence, and in a row or a column of the first frame or the second frame, is greater than or equal to a number; and determining the descriptor based on the luma information of the blocks in the first frame or the second frame other than the row or the column of the first frame or the second frame, in response to the determining that the number of pixels having the same pixel value at the same location in the frames of the first sequence or the second sequence, and in the row or the column of the first frame or the second frame, is greater than or equal to the number.

8 . The method of claim 1 , further comprising:
determining a descriptor of the first sequence based on the descriptor of the first frame; and determining a descriptor of the second sequence based on the descriptor of the second frame, wherein the determining the similarity comprises determining the similarity based on a comparison of the descriptor of the first sequence to the descriptor of the second sequence.

9 . The method of claim 8 , wherein the determining the descriptor of the first sequence comprises converting the descriptor of the first frame into the descriptor of the first sequence based on codes, and
the determining the descriptor of the second sequence comprises converting the descriptor of the second frame into the descriptor of the second sequence based on the codes.

10 . A non-transitory computer-readable recording storage medium storing a program to cause a computer to perform the method of claim 1 .

11 . A device for determining similarity between sequences, the device comprising:
a sequence obtainer configured to obtain a first sequence of frames and a second sequence of frames; a descriptor determiner configured to:
determine a descriptor of a first frame of the first sequence based on luma information of blocks in the first frame, the descriptor of the first frame comprising luma difference information that is determined based on the luma information of the blocks in the first frame; and
determine a descriptor of a second frame of the second sequence based on luma information of blocks in the second frame, the descriptor of the second frame comprising luma difference information that is determined based on the luma information of the blocks in the second frame; and
a similarity determiner configured to determine a similarity between the first sequence and the second sequence based on the descriptor of the first frame and the descriptor of the second frame.

12 . The device of claim 11 , wherein the similarity determiner is further configured to:
match the first frame with the second frame based on the descriptor of the first frame and the descriptor of the second frame; and determine a section in the second sequence that matches a section in the first sequence as a section similar to the section in the first sequence, based on the matched first frame and second frame.

13 . The device of claim 12 , wherein the similarity determiner is further configured to:
determine whether a Hamming distance between the descriptor of the first frame and the descriptor of the second frame is shortest; and match the first frame with the second frame in response to the similarity determiner determining that the Hamming distance between the descriptor of the first frame and the descriptor of the second frame is shortest.

14 . The device of claim 12 , wherein the similarity determiner is further configured to:
obtain a matching vector representing a matching relation between the matched first frame and second frame; determine a representative matching vector based on the obtained matching vector; and determine the section in the second sequence that matches the section in the first sequence as the section similar to the section in the first sequence, based on the representative matching vector.

15 . The device of claim 14 , wherein the similarity determiner is further configured to:
determine whether the matching vector of the first frame is not included in a range based on the representative matching vector; and re-match the first frame with a frame in the second sequence within the range based on the representative matching vector, in response to the similarity determiner determining that the matching vector of the first frame is not included in the range base

Assignees

• Samsung Electronics Co Ltd
• Univ Korea Res & Bus Found

Inventors

• Kwak Young-Jin
• Kim Kyung-Rae
• Kim Chang-Su
• Jang Won-Dong
• Lee Jun-Tae

Classifications

• G06V10/761

  Proximity, similarity or dissimilarity measures · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06V20/48 Primary

  Matching video sequences · CPC title

• G06V10/56

  relating to colour · CPC title

• G06V10/50

  by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis · CPC title