@inproceedings{65954c908b3644c79d7cadd3affb4d85,
title = "SAF: Semantic Attention Fusion Mechanism for Pedestrian Detection",
abstract = "Benefiting from deep learning methods, pedestrian detection has witnessed a great progress in recent years. However, many pedestrian detectors are prone to detect background instances, especially under urban scenes, which results in plenty of false positive detections. In this paper, we propose a semantic attention fusion mechanism (SAF) to increase the discriminability of detector. The SAF includes two key components, attention modules and reverse fusion blocks. Different from previous attention mechanisms which use attention modules for re-weighting the top features of network directly, the outputs of our attention modules are fused by reverse fusion blocks from high level layers to low level layers step by step, which aims at generating strong semantic features for pedestrian detections. Experiments on CityPersons dataset demonstrate the effectiveness of our SAF.",
keywords = "Background errors, Pedestrian detection, Semantic attention",
author = "Ruizhe Yu and Shunzhou Wang and Yao Lu and Huijun Di and Lin Zhang and Lihua Lu",
note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 16th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2019 ; Conference date: 26-08-2019 Through 30-08-2019",
year = "2019",
doi = "10.1007/978-3-030-29911-8_40",
language = "English",
isbn = "9783030299101",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "523--533",
editor = "Nayak, {Abhaya C.} and Alok Sharma",
booktitle = "PRICAI 2019",
address = "Germany",
}