Shypd.aiThe largest open index of AI toolsPointTinyBenchmark is an open-source toolbox for object localization and detection tasks. It provides benchmarks and algorithms for point-based and tiny object detection, built on mmdetection.
Trending
About
PointTinyBenchmark is an open-source toolbox designed for object localization and detection, with a specific focus on tiny objects and point-based methods. Built on top of mmdetection, it offers a comprehensive set of benchmarks and algorithms for researchers and developers in computer vision. The toolbox implements several key works, including 'Scale Match for TinyPerson Detection' (WACV2020), 'Detail Object Localization under Single Coarse Point Supervision' (CVPR2022), 'Point-to-Box Network for Accurate Object Detection via Single Point Supervision' (ECCV2022), and 'Spatial Self-Distillation for Object Detection with Inaccurate Bounding Boxes' (ICCV2023). It also anticipates 'CPR++: Object Localization via Single Coarse Point Supervision' (TPAMI2024). This tool is ideal for advancing research in computer vision, particularly for tasks involving small object detection and localization.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
