Search

EP-3884466-B1 - SYSTEMS AND METHODS FOR IMAGE SEGMENTATION

EP3884466B1EP 3884466 B1EP3884466 B1EP 3884466B1EP-3884466-B1

Inventors

  • WANG, FENG
  • MAO, LIJIAN
  • Sun, Haitao
  • XIONG, JIANPING

Dates

Publication Date
20260513
Application Date
20181225

Claims (12)

  1. A system for image segmentation, comprising: a storage medium storing a set of instructions; and at least one processor in communication with the storage medium, wherein when executing the set of instructions, the at least one processor is directed to: obtain an image; divide the image into a plurality of image blocks; determine, for each of the plurality of image blocks, a compressed representation of one or more features of the image block, the compressed representation of the one or more features of the image block including a set of binary codes, wherein the set of binary codes are generated according to a perceptual hash algorithm; group the plurality of image blocks into at least two different categories based on the number and distribution of "0" or "1" in each of the plurality of sets of binary codes, wherein if the number of "1" that exists in the last k bits of the respective set of binary codes is larger than a first threshold, the corresponding image block is designated as being of one category of the different categories, or if the number of "1" that exits in the last j bits of the respective set of binary codes is less than a second threshold, the corresponding image block is designated as being of another category of the different categories; and extract a region of interest from the image based on at least one of the at least two different categories of the plurality of image blocks.
  2. The system of claim 1, wherein for each of the plurality of image blocks, the one or more features of the image block include a texture feature of the image block.
  3. The system of claim 1 or claim 2, wherein the at least two different categories include a category associated with a foreground of the image and a category associated with a background of the image.
  4. The system of any one of claims 1-3, wherein to extract a region of interest from the image based on at least one of the at least two different categories of the plurality of image blocks, the at least one processor is directed to: determine, among the at least one of the at least two categories of image blocks, one or more target image blocks based on one or more neighboring image blocks thereof; and extract the region of interest from the image based on the one or more target image blocks.
  5. The system of claim 4, wherein to extract the region of interest from the image based on the one or more target image blocks, the at least one processor is directed to: merge the one or more target image blocks that belong to one of the at least two different categories-to form the region of interest.
  6. The system of claim 4, wherein to extract the region of interest from the image based on the one or more target image blocks, the at least one processor is directed to: connect two or more of the one or more target image blocks; and extract the region of interest from the image based on the connected two or more of the one or more target image blocks.
  7. The system of any one of claims 4-6, wherein to determine, among at least one of the at least two different categories of image blocks, one or more target image blocks based on one or more neighboring image blocks thereof, the at least one processor is directed to: determine a candidate image block and its one or more neighboring image blocks; determine, for each of the one or more neighboring image blocks of the candidate image block, a category to which the neighboring image block belongs; and designate the candidate image block as a target image block in response to that the one or more neighboring image blocks of the candidate image block belong to a same category.
  8. The system of any one of claims 1-7, wherein the at least one processor is further directed to: optimize the extracted region of interest to smoothen a boundary of the extracted region of interest.
  9. A method for image segmentation, implemented on a computing device including at least one processor and at least one storage medium, the method comprising: obtaining an image; dividing the image into a plurality of image blocks; determining, for each of the plurality of image blocks, a compressed representation of one or more features of the image block, the compressed representation of the one or more features of the image block including a set of binary codes, wherein the set of binary codes are generated according to a perceptual hash algorithm; grouping the plurality of image blocks into at least two different categories based on the number and distribution of "0" or "1" in each of the plurality of sets of binary codes, wherein if the number of "1" that exists in the last k bits of the respective set of binary codes is larger than a first threshold, the corresponding image block is designated as being of one category of the different categories, or if the number of "1" that exits in the last j bits of the respective set of binary codes is less than a second threshold, the corresponding image block is designated as being of another category of the different categories; and extracting a region of interest from the image based on at least one of the at least two different categories of the plurality of image blocks.
  10. The method of claim 9, wherein extracting a region of interest from the image based on at least one of the at least two different categories of the plurality of image blocks, including: determining, among at least one of the at least two different categories of image blocks, one or more target image blocks based on one or more neighboring image blocks thereof; and extracting the region of interest from the image based on the one or more target image blocks.
  11. The method of claim 10, wherein determining, among at least one of the at least two different categories of image blocks, one or more target image blocks based on one or more neighboring image blocks thereof, including: determining a candidate image block and its one or more neighboring image blocks; determining, for each of the one or more neighboring image blocks of the candidate image block, a category to which the neighboring image block belongs; and designating the candidate image block as a target image block in response to that the one or more neighboring image blocks of the candidate image block belong to a same category.
  12. A non-transitory computer readable medium, comprising a set of instructions for image segmentation, wherein when executed by at least one processor, the set of instructions directs the at least one processor to: obtain an image; divide the image into a plurality of image blocks; determine, for each of the plurality of image blocks, a compressed representation of one or more features of the image block, the compressed representation of the one or more features of the image block including a set of binary codes, wherein the set of binary codes are generated according to a perceptual hash algorithm; group the plurality of image blocks into at least two different categories based on the number and distribution of "0" or "1" in each of the plurality of sets of binary codes, wherein if the number of "1" that exists in the last k bits of the respective set of binary codes is larger than a first threshold, the corresponding image block is designated as being of one category of the different categories, or if the number of "1" that exits in the last j bits of the respective set of binary codes is less than a second threshold, the corresponding image block is designated as being of another category of the different categories; and extract a region of interest from the image based on at least one of the at least two different categories of the plurality of image blocks.

Description

TECHNICAL FIELD The present disclosure generally relates to methods, systems, and media for image processing. More particularly, the present disclosure relates to methods, systems, and media for segmenting an image. BACKGROUND With advances in image processing technology, image segmentation techniques have been widely used. A typical image segmentation technique may be used to segment different portions of an image. For example, an image segmentation technique can be used to segment out a main object (e.g., a foreground) from a background of the image. With such operation, the main object and the background can be processed separately. In some embodiments, the image segmentation technique may include an image editing operation that can assist in segregating the main object from the background of the image. For example, the image editing operation may permit a user to manually select a region in the image by, for example, manually tracing a boundary line around the main object. Similarly, the image editing operation may select a region in the image based on one or more user-selected points or areas that lie inside or outside the main object. Unfortunately, the image editing operation may have a number of shortcomings. For example, the image editing operation may be tedious and difficult to use for a user. Specifically, the manual input of boundary lines, points, and/or areas may cost a lot of time of the user. Moreover, the accuracy of the segmentation result may be largely affected by the input of the user. SUFYAN Y ABABNEH ET AL: "Automatic graph-cut based segmentation of bones from knee magnetic resonance images for osteoarthritis research", MEDICAL IMAGE ANALYSIS, OXFORD UNIVERSITY PRESS, OXOFRD, GB, vol. 15, no. 4, 31 January 2011 (2011-01-31), pages 438-448, XP028098914, ISSN: 1361-8415, DOI: 10.1016/J.MEDIA.2011.01.007 [retrieved on 2011-02-24] relates to a new, fully automated, content-based system for knee bone segmentation from magnetic resonance images (MRI). EP2863362A1 relates generally to image manipulation and segmentation and, more particularly, to scene segmentation from focal stack images. CHRISTOPH ZAUNER ED-ZAUNER: "Implementation and Benchmarking of Perceptual Image Hash Functions (thesis)", 20100701, 1 July 2010 (2010-07-01), pages i-94, XP002786627, Retrieved from the Internet: URL: https://www.phash.org/docs/pubs/thesis_zauner.pdf relates to a novel benchmarking framework, called Rihamark, for perceptual image hash functions. US9600746B2 relates to a technique for automatically detecting a salient region in an image. SUMMARY The invention is set out in the appended set of claims. Additional features will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following and the accompanying drawings or may be learned by production or operation of the examples. The features of the present disclosure may be realized and attained by practice or use of various aspects of the methodologies, instrumentalities and combinations set forth in the detailed examples discussed below. BRIEF DESCRIPTION OF THE DRAWINGS The disclosure will be understood more fully from the detailed description given below and from the accompanying drawings of various embodiments of the disclosure. The drawings, however, should not be taken to limit the disclosure to the specific embodiments, but are for explanation and understanding only. FIG. 1 is a schematic diagram illustrating an exemplary image processing system according to some embodiments of the present disclosure;FIG. 2 is a block diagram illustrating an exemplary computing device according to embodiments of the present disclosure of the present disclosure;FIG. 3 is a schematic diagram illustrating exemplary hardware and/or software components of an exemplary mobile device on which a terminal may be implemented according to some embodiments of the present disclosure;FIG. 4 is a block diagram illustrating an exemplary image processing device according to some embodiments of the present disclosure;FIG. 5 is a flowchart illustrating an exemplary process for segmenting an image according to some embodiments of the present disclosure;FIG. 6 is a flowchart illustrating an exemplary process for determine a fingerprint of each of the plurality of image blocks according to some embodiments of the present disclosure;FIG. 7 is a flowchart illustrating an exemplary process for extracting a region from an image according to some embodiments of the present disclosure;FIG. 8A illustrates an example of an image including a region of interest according to some embodiments of the present disclosure;FIG. 8B illustrates examples of the image blocks of the image according to some embodiments of the present disclosure;FIG. 9A illustrates five examples of neighboring image blocks of a candidate image block according to some embodiments of the present disclosure;FIG. 9B illustrates the target image b