Show HN: I generated a "stress test" of 200 rare defects from 7 real photos
Summary
The post describes a pipeline that generates 200 rare defect images from 7 real samples to benchmark model recall in vision systems. It releases a CC0 dataset of broken insulators with labeled images for validation of rare failure modes and discusses the importance of edge-case testing. The discussion emphasizes synthetic data as a tool for evaluating recall on low-probability events and invites feedback on validation approaches.