Automatic image segmentation with superpixels and image-level labels

Xinlin Xie, Gang Xie, Xinying Xu, Lei Cui, Jinchang Ren

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Automatically and ideally segmenting the semantic region of each object in an image will greatly improve the precision and efficiency of subsequent image processing. We propose an automatic image segmentation algorithm based on superpixels and image-level labels. The proposed algorithm consists of three stages. At the stage of superpixel segmentation, we adaptively generate the initial number of superpixels using the minimum spatial distance and the total number of pixels in the image. At the stage of superpixel merging, we define small superpixels and directly merge the most similar superpixel pairs without considering the adjacency, until the number of superpixels equals the number of groupings contained in image-level labels. Furthermore, we add a stage of reclassification of disconnected regions after superpixel merging to enhance the connectivity of segmented regions. On the widely used Microsoft Research Cambridge data set and Berkeley segmentation data set, we demonstrate that our algorithm can produce high-precision image segmentation results compared with the state-of-the-art algorithms.
LanguageEnglish
Article number8607985
Pages10999-11009
Number of pages11
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 10 Jan 2019

Fingerprint

Image segmentation
Labels
Merging
Image processing
Pixels
Semantics

Keywords

  • image segmentation
  • superpixels
  • image-level labels
  • disconnected regions

Cite this

Xie, Xinlin ; Xie, Gang ; Xu, Xinying ; Cui, Lei ; Ren, Jinchang. / Automatic image segmentation with superpixels and image-level labels. In: IEEE Access. 2019 ; Vol. 7. pp. 10999-11009.
@article{8e1605ae6de440c4a2e100acd7527707,
title = "Automatic image segmentation with superpixels and image-level labels",
abstract = "Automatically and ideally segmenting the semantic region of each object in an image will greatly improve the precision and efficiency of subsequent image processing. We propose an automatic image segmentation algorithm based on superpixels and image-level labels. The proposed algorithm consists of three stages. At the stage of superpixel segmentation, we adaptively generate the initial number of superpixels using the minimum spatial distance and the total number of pixels in the image. At the stage of superpixel merging, we define small superpixels and directly merge the most similar superpixel pairs without considering the adjacency, until the number of superpixels equals the number of groupings contained in image-level labels. Furthermore, we add a stage of reclassification of disconnected regions after superpixel merging to enhance the connectivity of segmented regions. On the widely used Microsoft Research Cambridge data set and Berkeley segmentation data set, we demonstrate that our algorithm can produce high-precision image segmentation results compared with the state-of-the-art algorithms.",
keywords = "image segmentation, superpixels, image-level labels, disconnected regions",
author = "Xinlin Xie and Gang Xie and Xinying Xu and Lei Cui and Jinchang Ren",
note = "{\circledC} 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission",
year = "2019",
month = "1",
day = "10",
doi = "10.1109/ACCESS.2019.2891941",
language = "English",
volume = "7",
pages = "10999--11009",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE",

}

Automatic image segmentation with superpixels and image-level labels. / Xie, Xinlin; Xie, Gang; Xu, Xinying; Cui, Lei; Ren, Jinchang.

In: IEEE Access, Vol. 7, 8607985, 10.01.2019, p. 10999-11009.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Automatic image segmentation with superpixels and image-level labels

AU - Xie, Xinlin

AU - Xie, Gang

AU - Xu, Xinying

AU - Cui, Lei

AU - Ren, Jinchang

N1 - © 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission

PY - 2019/1/10

Y1 - 2019/1/10

N2 - Automatically and ideally segmenting the semantic region of each object in an image will greatly improve the precision and efficiency of subsequent image processing. We propose an automatic image segmentation algorithm based on superpixels and image-level labels. The proposed algorithm consists of three stages. At the stage of superpixel segmentation, we adaptively generate the initial number of superpixels using the minimum spatial distance and the total number of pixels in the image. At the stage of superpixel merging, we define small superpixels and directly merge the most similar superpixel pairs without considering the adjacency, until the number of superpixels equals the number of groupings contained in image-level labels. Furthermore, we add a stage of reclassification of disconnected regions after superpixel merging to enhance the connectivity of segmented regions. On the widely used Microsoft Research Cambridge data set and Berkeley segmentation data set, we demonstrate that our algorithm can produce high-precision image segmentation results compared with the state-of-the-art algorithms.

AB - Automatically and ideally segmenting the semantic region of each object in an image will greatly improve the precision and efficiency of subsequent image processing. We propose an automatic image segmentation algorithm based on superpixels and image-level labels. The proposed algorithm consists of three stages. At the stage of superpixel segmentation, we adaptively generate the initial number of superpixels using the minimum spatial distance and the total number of pixels in the image. At the stage of superpixel merging, we define small superpixels and directly merge the most similar superpixel pairs without considering the adjacency, until the number of superpixels equals the number of groupings contained in image-level labels. Furthermore, we add a stage of reclassification of disconnected regions after superpixel merging to enhance the connectivity of segmented regions. On the widely used Microsoft Research Cambridge data set and Berkeley segmentation data set, we demonstrate that our algorithm can produce high-precision image segmentation results compared with the state-of-the-art algorithms.

KW - image segmentation

KW - superpixels

KW - image-level labels

KW - disconnected regions

UR - https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6287639

U2 - 10.1109/ACCESS.2019.2891941

DO - 10.1109/ACCESS.2019.2891941

M3 - Article

VL - 7

SP - 10999

EP - 11009

JO - IEEE Access

T2 - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8607985

ER -