Probiou on DeepLearning.Earthhttps://deeplearning.earth/tags/probiou/Recent content in Probiou on DeepLearning.EarthHugo -- gohugo.ioen-usFri, 10 Jul 2026 09:00:00 +0700Rotated Faster R-CNN on DOTA without custom CUDA: sampled rIoU, ProbIoU, and an 83.4% checkpointhttps://deeplearning.earth/posts/2026-07-10_rotated_faster_rcnn_probiou_dota/Fri, 10 Jul 2026 09:00:00 +0700https://deeplearning.earth/posts/2026-07-10_rotated_faster_rcnn_probiou_dota/Satellite scene with oriented bounding boxes Oriented object detection on satellite imagery lives or dies on rotated IoU — the overlap between two arbitrarily angled rectangles. Frameworks like MMRotate ship exact, CUDA-accelerated rotated IoU for training and inference. That is fast and precise, but it ties you to a heavy stack: MMCV custom ops, version pins, and compiled extensions that are painful to ship in a lean research codebase. OrientedDet takes a different path for v1: a full Python / PyTorch detector with no custom CUDA kernels and no MMCV runtime dependency.