Technology Category
Computer Vision: Object Detection & CategorizationSearch and Optimization: Learning to SearchMachine Learning: Transfer, Domain Adaptation, Multi-Task Learning
Adaptive image zoom-in with bounding box transformation for UAV object detection
🤖 AI Summary
This work addresses the challenge of detecting small and sparsely distributed objects in drone-captured images, which often limits the performance of general-purpose detectors. To this end, the authors propose an adaptive non-uniform scaling framework that learns to locally magnify foreground regions through a lightweight offset prediction mechanism, coupled with a corner-aligned spatial transformation of bounding boxes. This approach enables end-to-end, architecture-agnostic feature enhancement and detection. By introducing a bounding-box-based scaling objective function, the method achieves significant performance gains on the VisDrone, UAVDT, and SeaDronesSee benchmarks, with an improvement of over 8.4 mAP on SeaDronesSee while adding only approximately 3 ms to inference latency.
Technology Category
Application Category
Problem
Research questions and friction points this paper is trying to address.
UAV object detection
small object detection
image zoom-in
bounding box transformation
adaptive zooming
Innovation
Methods, ideas, or system contributions that make the work stand out.
adaptive zoom-in
bounding box transformation
UAV object detection
non-uniform image warping
small object detection
🔎 Similar Papers
No similar papers found.
