R-CNN for Small Object Detection

By Chenyi Chen, Ming-Yu Liu, Oncel Tuzel, Jianxiong Xiao. Asian Conference on Computer Vision, 2016

397 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4bbb326375d17fc10892c1b6aa684db4af6f66bd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Junwei Han
2. Gong Cheng
3. Xiwen Yao
4. Zehao Huang
5. Chenhongyi Yang
6. Ke Yan
7. Xiang Yuan
8. Qinghua Zeng

Citations in the corpus

• FSSD: Feature Fusion Single Shot Multibox Detector (2017, background): “Besides, small objects also rely on the context information heavily [2].”
• A Survey of the Four Pillars for Small Object Detection: Multiscale Representation, Contextual Information, Super-Resolution, and Region Proposal (2022, methodology): “ContextNet [53]: Augmented R-CNN [53] could be considered as the first detector focusing on small object detection.”
• Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, methodology): “Chen et al. [28] employed the representations of context regions which encompass the proposal patches for subsequent recognition.”
• SuperYOLO: Super Resolution Assisted Object Detection in Multimodal Remote Sensing Imagery (2022, background): “In the recent literature, the performance of small object detection can be improved by multiscale feature learning [29], [30] and context-based detection [31].”
• QueryDet: Cascaded Sparse Query for Accelerating High-Resolution Small Object Detection (2021, methodology): “These methods can be mainly categorized into four types: 1) increasing the resolution of input features [31, 2, 23, 11, 28, 46, 26, 53]; 2) oversampling and strong data augmentation [31, 21, 75]; 3) incorporating context information [7, 64, 5], and 4) scale-aware training [28, 47, 48, 27].”
• Object Detection in 20 Years: A Survey (2019, neutral): “Recent deep learning-based detectors can also be improved with local context by simply enlarging the networks’ receptive field or the size of object proposals [79], [80], [81], [82], [83], [84], [97].”