Bootstrap Your Object Detector via Mixed Training
Dec 6, 2021
Speakers
Zheng Zhang
Speaker · 0 followers
Fangyun Wei
Speaker · 0 followers
Yutong Lin
Speaker · 0 followers
Stephen Lin
Speaker · 0 followers
About
We introduce MixTraining, a new training paradigm for object detection that can improve the performance of existing detectors for free. MixTraining enhances data augmentation by utilizing augmentations of different strengths while excluding the strong augmentations of certain training samples that may be detrimental to training. In addition, it addresses localization noise and missing labels in human annotations by incorporating pseudo boxes that can compensate for these errors. Both of these MixTraining capabilities are made possible through bootstrapping on the detector, which can be used to predict the difficulty of training on a strong augmentation, as well as to generate reliable pseudo boxes thanks to the robustness of neural networks to labeling error. MixTraining is found to bring consistent improvements across various detectors on the COCO dataset. In particular, the performance of Faster R-CNN <cit.> with a ResNet-50 <cit.> backbone is improved from 41.7 mAP to 44.0 mAP, and the accuracy of Cascade-RCNN <cit.> with a Swin-Small <cit.> backbone is raised from 50.9 mAP to 52.8 mAP.We introduce MixTraining, a new training paradigm for object detection that can improve the performance of existing detectors for free. MixTraining enhances data augmentation by utilizing augmentations of different strengths while excluding the strong augmentations of certain training samples that may be detrimental to training. In addition, it addresses localization noise and missing labels in human annotations by incorporating pseudo boxes that can compensate for these errors. Both of these M…
Organizer
NeurIPS 2021
Account · 1.9k followers
About NeurIPS 2021
Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Oral Session 5: Vision Applications
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Partial-Attribution Instance Segmentation for Astronomical Source Detection and Deblending
Watch later
Ryan Hausen, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Compositional Reinforcement Learning from Logical Specifications
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Q&A for Raquel Urtasun and Olga Russakovsky
Watch later
Serena Yeung, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
MixACM: Mixup-Based Robustness Transfer via Distillation of Activated Channel Maps
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization
Watch later
Jiaqi Gu, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%