TY - GEN
T1 - Self-Compensating Learning for Few-Shot Segmentation
AU - Wang, Jin
AU - Zhang, Bingfeng
AU - Liu, Weifeng
AU - Liu, Baodi
AU - Yu, Siyue
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Few-shot segmentation (FSS) has witnessed rapid development. Most existing approaches extract prototypes from support images to segment query images. However, the integrity and validity of these support prototypes cannot be guaranteed. To solve the above drawbacks, we propose a self-compensating strategy, aiming to provide query-aware support information, to build more effective matching between support information and query images. Specifically, we design a prototype compensating module to mine useful information from the query prediction, to update original support prototypes as new query-aware support prototypes. Then the updated prototypes are utilized to perform the second matching with query features. In addition, we also compensate the information of original prior masks on the second matching phase, to improve the quality of prior masks. With improved prototype representations and prior knowledge, our approach can directly improve the performance of different approaches with new state-of-the-art performances.
AB - Few-shot segmentation (FSS) has witnessed rapid development. Most existing approaches extract prototypes from support images to segment query images. However, the integrity and validity of these support prototypes cannot be guaranteed. To solve the above drawbacks, we propose a self-compensating strategy, aiming to provide query-aware support information, to build more effective matching between support information and query images. Specifically, we design a prototype compensating module to mine useful information from the query prediction, to update original support prototypes as new query-aware support prototypes. Then the updated prototypes are utilized to perform the second matching with query features. In addition, we also compensate the information of original prior masks on the second matching phase, to improve the quality of prior masks. With improved prototype representations and prior knowledge, our approach can directly improve the performance of different approaches with new state-of-the-art performances.
KW - Few-shot segmentation
KW - prior mask
KW - prototype
KW - self-compensation
UR - http://www.scopus.com/inward/record.url?scp=85180788494&partnerID=8YFLogxK
U2 - 10.1109/ICIP49359.2023.10223162
DO - 10.1109/ICIP49359.2023.10223162
M3 - Conference Proceeding
AN - SCOPUS:85180788494
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 1320
EP - 1324
BT - 2023 IEEE International Conference on Image Processing, ICIP 2023 - Proceedings
PB - IEEE Computer Society
T2 - 30th IEEE International Conference on Image Processing, ICIP 2023
Y2 - 8 October 2023 through 11 October 2023
ER -