TY - JOUR
T1 - Digital image semantic segmentation algorithms
T2 - A survey
AU - Song, Wei
AU - Zheng, Nan
AU - Zheng, Rui
AU - Zhao, Xiao Bing
AU - Wang, Antai
N1 - Funding Information:
The financial support and general support for this survey provided by the National Natural Science Foundation of China (No. 61503424, 61331013),National Scholarship Foundation of China, Promotion plan for young teachers’ scientific research ability of Minzu University of China, Master’s independent research project of Minzu university of China (182164) and the MUC 111 Project are gratefully acknowledged.
Publisher Copyright:
© 2019, Ubiquitous International. All rights reserved.
PY - 2019/1
Y1 - 2019/1
N2 - In the field of computer vision, image semantic segmentation is an important research branch and it is also a challenging task. Applications such as autonomous driving, Unmanned Aerial Vehicle System (UAVS), and even virtual or augmented reality systems require accurate and efficient segmentation mechanisms. With the rise of deep learning methods, image semantic segmentation is more and more concerned by relevant researchers. In order to understand the research status, existing problems and development prospects of image semantic segmentation, this paper introduces the mainstream image semantic segmentation methods on the basis of extensive survey. First of all, we introduce the background concept of image semantic segmentation, generalize the commonly used image semantic segmentation methods, and compare the segmentation results of each method. After that, the commonly used image semantic segmentation datasets are summarized. At the same time, several commonly evaluation standards are introduced. Finally, the future development trend of image semantic segmentation is prospected, with a view to providing some ideas for researchers who wish to engage in this field.
AB - In the field of computer vision, image semantic segmentation is an important research branch and it is also a challenging task. Applications such as autonomous driving, Unmanned Aerial Vehicle System (UAVS), and even virtual or augmented reality systems require accurate and efficient segmentation mechanisms. With the rise of deep learning methods, image semantic segmentation is more and more concerned by relevant researchers. In order to understand the research status, existing problems and development prospects of image semantic segmentation, this paper introduces the mainstream image semantic segmentation methods on the basis of extensive survey. First of all, we introduce the background concept of image semantic segmentation, generalize the commonly used image semantic segmentation methods, and compare the segmentation results of each method. After that, the commonly used image semantic segmentation datasets are summarized. At the same time, several commonly evaluation standards are introduced. Finally, the future development trend of image semantic segmentation is prospected, with a view to providing some ideas for researchers who wish to engage in this field.
KW - Deep learning
KW - Image semantic segmentation
KW - Neural network
UR - http://www.scopus.com/inward/record.url?scp=85059375949&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85059375949&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:85059375949
SN - 2073-4212
VL - 10
SP - 196
EP - 211
JO - Journal of Information Hiding and Multimedia Signal Processing
JF - Journal of Information Hiding and Multimedia Signal Processing
IS - 1
ER -