首页|Relative difficulty distillation for semantic segmentation

Relative difficulty distillation for semantic segmentation

扫码查看
Current knowledge distillation(KD)methods primarily focus on transferring various structured knowledge and designing corresponding optimization goals to encourage the student network to imitate the output of the teacher network.However,introducing too many additional optimization objectives may lead to unstable training,such as gradient conflicts.Moreover,these methods ignored the guidelines of relative learning difficulty between the teacher and student networks.Inspired by human cognitive science,in this paper,we redefine knowledge from a new perspective-the student and teacher networks'relative difficulty of samples,and propose a pixel-level KD paradigm for semantic segmentation named relative difficulty distillation(RDD).We propose a two-stage RDD framework:teacher-full evaluated RDD(TFE-RDD)and teacher-student evaluated RDD(TSE-RDD).RDD allows the teacher network to provide effective guidance on learning focus without additional optimization goals,thus avoiding adjusting learning weights for multiple losses.Extensive experimental evaluations using a general distillation loss function on popular datasets such as Cityscapes,CamVid,Pascal VOC,and ADE20k demonstrate the effectiveness of RDD against state-of-the-art KD methods.Additionally,our research showcases that RDD can integrate with existing KD methods to improve their upper performance bound.Codes are available at https://github.com/sunyueue/RDD.git.

knowledge distillationsemantic segmentationrelative difficultysample weightingprediction discrepancy

Dong LIANG、Yue SUN、Yun DU、Songcan CHEN、Sheng-Jun HUANG

展开 >

MHT Key Laboratory of Pattern Analysis and Machine Intelligence,College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China

Shenzhen Research Institute,Nanjing University of Aeronautics and Astronautics,Shenzhen 518000,China

National Natural Science Foundation of ChinaNational Natural Science Foundation of ChinaNational Natural Science Foundation of ChinaNational Key R&D Program of ChinaNatural Science Foundation of Jiangsu ProvinceNatural Science Foundation of Jiangsu ProvinceShenzhen Science and Technology Program

6227222962076124622226052020AAA0107000BK20222012BK20211517JCYJ2023080714200-1004

2024

中国科学:信息科学(英文版)
中国科学院

中国科学:信息科学(英文版)

CSTPCDEI
影响因子:0.715
ISSN:1674-733X
年,卷(期):2024.67(9)