Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces ROI-Gated SAHI, an inference-time framework that enhances small object detection in high-resolution images by using a lightweight proposer to focus on foreground regions, thereby reducing computational waste on background tiles. Evaluated on the COCO128 dataset, the static ROI-gating method showed a speed ratio of 0.88 compared to Full SAHI, with a lower mAP@0.5 of 0.6602, while an adaptive routing policy improved mean latency and achieved a speedup of 1.02×. In a sparse-to-dense case study, ROI-gating demonstrated significant speedups, ranging from 0.96× to 6.90×, highlighting its effectiveness in optimizing inference for sparse scenes.
ROI-Gated SAHI can achieve up to 6.90× speedups in object detection by intelligently focusing on informative areas, revolutionizing efficiency in high-resolution image processing.
Slicing-Aided Hyper Inference (SAHI) improves small object detection in high-resolution images but often spends substantial compute on background tiles. We propose region-of-interest (ROI)-Gated SAHI, an inference-time framework that introduces a lightweight proposer to localize foreground regions and restrict sliced refinement to informative areas. We evaluate the framework in two settings. On the COCO128 full split dataset comprising 128 images, static ROI-gating is slower on average than Full SAHI, achieving a speed ratio of 0.88, and yields a lower mAP@0.5 of 0.6602 compared with 0.7569 for Full SAHI. A simple adaptive routing policy with $\tau =$ 0.4 educes the mean latency, achieving a slight gain of 1.02$\times$ over Full SAHI. On a three-image sparse-to-dense case study, ROI-gating achieves speedups ranging from 0.96$\times$ to 6.90$\times$ with a mean speedup of 3.41$\times$. These results show that ROI-gating is most beneficial in sparse scenes and requires policy-based routing for robust average behavior.