Small object detection using sparsely-connected convolution for less compute with TinyDet
Small object detection using sparsely-connected convolution for less compute with TinyDet
TinyDet: Accurate Small Object Detection in Lightweight Generic Detectors
arXiv paper abstract https://arxiv.org/abs/2304.03428
arXiv PDF paper https://arxiv.org/pdf/2304.03428.pdf
Small object detection requires … scan a large number of positions on image feature maps, which is … hard for computation- and energy-efficient …
To … detect … with limited computation, … propose a two-stage lightweight detection framework with extremely low computation complexity, termed as TinyDet.
It enables high-resolution feature maps for dense anchoring to better cover small objects, proposes a sparsely-connected convolution for computation reduction, enhances the early stage features in the backbone, and addresses the feature misalignment problem for accurate small object detection.
… TinyDet-M … is the first detector that has an AP over 30 with less than 1 GFLOPs …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
