Learning Data Augmentation Strategies for Object Detection
lastmod 2026-11-10created 2019-11-073 edits
Learning Data Augmentation Strategies for Object Detection is a paper that proposes finding data augmentation policies for training object detection models through search rather than designing them by hand. It defines a space of augmentation policies combining color transformations, geometric transformations, and bounding-box-level transformations, finds the optimal policy with reinforcement learning, and shows that it can be applied to other datasets and models.
Documentation
- Learning Data Augmentation Strategies for Object Detection
- https://arxiv.org/abs/1906.11172v1