arrow
返回

Rethinking Self-Attention for Multispectral Object Detection

delete2024-11-01
delete0
delete
OA
AI
S
Sijie Hu *
F
Fabien Bonardi
S
Samia Bouchafa
H
Helmut Prendinger
D
Désiré Sidibé
DOI:10.1109/TITS.2024.3412417delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Data from different modalities, such as infrared and visible images, can offer complementary information, and integrating such information can significantly enhance the capabilities of a system to perceive and recognize its surroundings. Thus, multi-modal object detection has widespread applications, particularly in challenging weather conditions like low-light scenarios. The core of multi-modal fusion lies in developing a reasonable fusion strategy, which can fully exploit the complementary features of different modalities while preventing a significant increase in model complexity. To this end, this paper proposes a novel lightweight cross-fusion module named Channel-Patch Cross Fusion (CPCF), which leverages Channel-wise Cross-Attention (CCA), Patch-wise Cross-Attention (PCA) and Adaptive Gating (AG) to encourage mutual rectification among different modalities. This process simultaneously explores commonalities across modalities while maintaining the uniqueness of each modality. Furthermore, we design a versatile intermediate fusion framework that can leverage CPCF to enhance the performance of multi-modal object detection. The proposed method is extensively evaluated on multiple public multi-modal datasets, namely FLIR, LLVIP, and DroneVehicle. The experiments indicate that our method yields consistent performance gains across various benchmarks and can be extended to different types of detectors, further demonstrating its robustness and generalizability. Our codes are available at https://github.com/Superjie13/CPCF_Multispectral.
Keyword:
Multispectral
attention
intermediate fusion
object detection
deep learning

期刊

IEEE Transactions on Intelligent Transportation Systems 封面图
IEEE Transactions on Intelligent Transportation Systems
IF:
8.4
论文数:
9.5K
被引数:
6.3W

机构

R
research organization of information & systems (rois)
学者数:
2.8K
论文数: 3.2K
被引数: 2
U
Universite Paris Saclay
学者数:
7.3W
论文数: 5.3W
被引数: 540
引用论文

引用论文

Role of Gamma Knife surgery in the management of pineal region tumors
err2007-12-01
err0
errOAAI
errGregory P. Lekovic; L. Fernando Gonzalez; Andrew G. Shetter; Randall W. Porter; Kris A. Smith; David Brachman; Robert F. Spetzler
err分享
err收藏
Selective Search for Object Recognition
err2013-04-02
err3.9K
PREAI
errUijlings, J. R. R.; van de Sande, K. E. A.; Gevers, T.; Smeulders, A. W. M.
err分享
err收藏
err分享
err收藏
Immunohistochemical expression of HER1, HER3, and HER4 in HER2‐positive breast cancer patients treated with trastuzumab‐containing neoadjuvant chemotherapy
err2010-01-19
err0
errOAAI
errKan Yonemori; Koji Tsuta; Chikako Shimizu; Yutaka Hatanaka; Akihiro Hirakawa; Makiko Ono; Tsutomu Kouno; Noriyuki Katsumata; Masashi Ando; Kenji Tamura; Tadashi Hasegawa; Takayuki Kinoshita; Yasuhiro Fujiwara
err分享
err收藏
Cross‐nucleation of polybutene‐1 Form II on Form I seeds with different morphology
err2020-02-05
err0
errOAAI
errWei Wang; Bao Wang; Enrico Carmeli; Zefan Wang; Zhe Ma; Dario Cavallo
err分享
err收藏
学者 查看更多内容