Scale Match for Tiny Person Detection

By Yuqi Gong, Nan Jiang, Qixiang Ye, Zhenjun Han. IEEE Workshop/Winter Conference on Applications of Computer Vision, 2019

349 citations recorded by the source corpus.

Visual object detection has achieved unprecedented advance with the rise of deep convolutional neural networks. However, detecting tiny objects (for example tiny persons less than 20 pixels) in large-scale images remains not well investigated. The extremely small objects raise a grand challenge about feature representation while the massive and complex backgrounds aggregate the risk of false alarms. In this paper, we introduce a new benchmark, referred to as TinyPerson, opening up a promising direction for tiny object detection in a long distance and with massive backgrounds. We experimentally find that the scale mismatch between the dataset for network pre-training and the dataset for detector learning could deteriorate the feature representation and the detectors. Accordingly, we propose a simple yet effective Scale Match approach to align the object scales between the two datasets for favorable tiny-object representation. Experiments show the significant performance gain of our proposed approach over state-of-the-art detectors, and the challenging aspects of TinyPerson related to real-world scenarios. The TinyPerson benchmark and the code for our approach will be publicly available1.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/67422d00c9e8293e94cab9be57f52f3cdb7ac03a

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. Gong Cheng
2. Junwei Han
3. Xiwen Yao
4. Guisong Xia
5. Wen Yang
6. Ke Yan
7. Xiang Yuan
8. Qinghua Zeng

Citations in the corpus

• Attention-Guided Pyramid Context Networks for Detecting Infrared Small Target Under Complex Background (2023, background): “Infrared Small Target Detection Networks: In visible image data, small target detection is gradually receiving more attention [30], [31], [32], [33], [34].”
• Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, methodology): “In this task, the terms tiny and small are typically defined by an area threshold [6] or length threshold [7], [8].”
• A Normalized Gaussian Wasserstein Distance for Tiny Object Detection (2021, neutral): “While tiny objects (less than 16× 16 pixels in the AI-TOD dataset [29]) often exhibit with extremely limited appearance information, which increases difficulty in learning discriminative features, leading to enormous failure cases when detecting tiny objects [25, 29, 35].”
• Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark (2022, neutral): “Existing tiny object detection benchmarks include TinyPerson benchmark (Yu et al., 2020) and AI-TOD benchmark (Wang et al.”
• Tiny Object Detection in Aerial Images (2021, methodology): “Specifically, absolute size Sa(·) and relative size Sr(·) of an object can be calculated by [6]:”