Summary of Dualteacher: Bridging Coexistence Of Unlabelled Classes For Semi-supervised Incremental Object Detection, by Ziqi Yuan et al.
DualTeacher: Bridging Coexistence of Unlabelled Classes for Semi-supervised Incremental Object Detection
by Ziqi Yuan, Liyuan Wang, Wenbo Ding, Xingxing Zhang, Jiachen Zhong, Jianyong Ai, Jianmin Li, Jun Zhu
First submitted to arxiv on: 13 Dec 2023
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A semi-supervised incremental object detection (SSIOD) approach is proposed to accommodate new classes effectively in real-world applications. The traditional supervised IOD assumes fully annotated data, which is rare and expensive. Instead, SSIOD considers a setting where the detector learns from a few labelled data and massive unlabelled data without forgetting old classes. A DualTeacher method is introduced, using two teacher models for old and new classes to instruct the student, achieving superior performance with limited resource overhead. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In this paper, researchers tackle a common problem in object detection: teaching machines to recognize objects they haven’t seen before. This is important because, in real life, we encounter many new things that we don’t know how to label yet. The usual way of doing this (supervised learning) isn’t practical because it requires lots of labeled data, which can be hard to get. So, the researchers came up with a new approach called semi-supervised incremental object detection (SSIOD). It’s like trying to teach a child to recognize different animals without labeling each one individually. The approach uses two “teacher” models that help the student model learn what to do. This method outperforms others in benchmarks and could be useful for applications like self-driving cars. |
Keywords
» Artificial intelligence » Object detection » Semi supervised » Student model » Supervised
