Saved in:
Bibliographic Details
Main Authors: Deng, Jieren, Zhou, Xin, Tian, Hao, Pan, Zhihong, Aguiar, Derek
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2303.05015
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911761015242752
author Deng, Jieren
Zhou, Xin
Tian, Hao
Pan, Zhihong
Aguiar, Derek
author_facet Deng, Jieren
Zhou, Xin
Tian, Hao
Pan, Zhihong
Aguiar, Derek
contents Distilling the structured information captured in feature maps has contributed to improved results for object detection tasks, but requires careful selection of baseline architectures and substantial pre-training. Self-distillation addresses these limitations and has recently achieved state-of-the-art performance for object detection despite making several simplifying architectural assumptions. Building on this work, we propose Smooth and Stepwise Self-Distillation (SSSD) for object detection. Our SSSD architecture forms an implicit teacher from object labels and a feature pyramid network backbone to distill label-annotated feature maps using Jensen-Shannon distance, which is smoother than distillation losses used in prior work. We additionally add a distillation coefficient that is adaptively configured based on the learning rate. We extensively benchmark SSSD against a baseline and two state-of-the-art object detector architectures on the COCO dataset by varying the coefficients and backbone and detector networks. We demonstrate that SSSD achieves higher average precision in most experimental settings, is robust to a wide range of coefficients, and benefits from our stepwise distillation procedure.
format Preprint
id arxiv_https___arxiv_org_abs_2303_05015
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Smooth and Stepwise Self-Distillation for Object Detection
Deng, Jieren
Zhou, Xin
Tian, Hao
Pan, Zhihong
Aguiar, Derek
Computer Vision and Pattern Recognition
Distilling the structured information captured in feature maps has contributed to improved results for object detection tasks, but requires careful selection of baseline architectures and substantial pre-training. Self-distillation addresses these limitations and has recently achieved state-of-the-art performance for object detection despite making several simplifying architectural assumptions. Building on this work, we propose Smooth and Stepwise Self-Distillation (SSSD) for object detection. Our SSSD architecture forms an implicit teacher from object labels and a feature pyramid network backbone to distill label-annotated feature maps using Jensen-Shannon distance, which is smoother than distillation losses used in prior work. We additionally add a distillation coefficient that is adaptively configured based on the learning rate. We extensively benchmark SSSD against a baseline and two state-of-the-art object detector architectures on the COCO dataset by varying the coefficients and backbone and detector networks. We demonstrate that SSSD achieves higher average precision in most experimental settings, is robust to a wide range of coefficients, and benefits from our stepwise distillation procedure.
title Smooth and Stepwise Self-Distillation for Object Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2303.05015