Proceedings of the AAAI Conference on Artificial Intelligence · 2022 · 88 citations · 30 references
ComputingComputer ScienceEducational TechnologyComputer VisionMachine LearningMachine VisionDeep LearningFeature DetectionObject DetectionObject RecognitionPattern RecognitionRepresentation LearningCompressionKnowledge DistillationModel CompressionOne-stage DetectorsPrecision
Knowledge Distillation (KD) is a widely-used technology to inherit information from cumbersome teacher models to compact student models, consequently realizing model compression and acceleration. Compared with image classification, object detection is a more complex task, and designing specific KD methods for object detection is non-trivial. In this work, we elaborately study the behaviour difference between the teacher and student detection models, and obtain two intriguing observations: First, the teacher and student rank their detected candidate boxes quite differently, which results in their precision discrepancy. Second, there is a considerable gap between the feature response differences and prediction differences between teacher and student, indicating that equally imitating all the feature maps of the teacher is the sub-optimal choice for improving the student's accuracy. Based on the two observations, we propose Rank Mimicking (RM) and Prediction-guided Feature Imitation (PFI) for distilling one-stage detectors, respectively. RM takes the rank of candidate boxes from teachers as a new form of knowledge to distill, which consistently outperforms the traditional soft label distillation. PFI attempts to correlate feature differences with prediction differences, making feature imitation directly help to improve the student's accuracy. On MS COCO and PASCAL VOC benchmarks, extensive experiments are conducted on various detectors with different backbones to validate the effectiveness of our method. Specifically, RetinaNet with ResNet50 achieves 40.4% mAP on MS COCO, which is 3.5% higher than its baseline, and also outperforms previous KD methods.
30
Deep Residual Learning for Image Recognition
Kaiming He, Xiangyu Zhang, Shaoqing Ren et al. · 2016 · 222.1K citations · Full text
Feature Pyramid Networks for Object Detection
Tsung-Yi Lin, Piotr Dollár, Ross Girshick et al. · 2017 · 28.7K citations
MobileNetV2: Inverted Residuals and Linear Bottlenecks
Mark Sandler, Andrew Howard, Menglong Zhu et al. · 2018 · 25.4K citations
The Pascal Visual Object Classes (VOC) Challenge
Mark Everingham, Luc Van Gool, Christopher K. I. Williams et al. · International Journal of Computer Vision · 2009 · 19.5K citations · Full text