Deep Learning-Based Object Tracking System By Using Visual and Thermal Infrared Fusion

Date

2023-8-28

Author

Türkoğlu, Abbas

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

410
views

653
downloads

Cite This

Object tracking in computer vision presents a formidable challenge, particularly when faced with adverse conditions like occlusion, variations in illumination, and motion blur. In recent years, deep learning has shown great promise for object tracking. However, the vast majority of deep learning-based object trackers use only visible band images. This limits their performance in challenging conditions, as thermal infrared electromagnetic waves can provide additional information about the object, such as its temperature. This thesis proposes a deep learning-based object tracking system that uses visual band (RGB) and thermal infrared (RGBT) fused images. The system consists of two main components: a feature extractor and a tracker. The feature extractor extracts features from both RGB and thermal infrared (TIR) images. The tracker then uses these features to track the object in the next frame. The proposed system is evaluated on the RGBT234 and LasHeR datasets, which are the mostly used RGBT object tracking datasets in the literature. The results show that the proposed system outperforms state-of-the-art RGB object trackers on the RGBT234 and LasHeR datasets.

Subject Keywords

Deep learning, object tracking, RGBT object tracking, computer vision, thermal and visual fusion

URI

https://hdl.handle.net/11511/105292

Collections

Graduate School of Informatics, Thesis