Reproduce “Segment Anything in Defect Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Bozhen Hu, Bin Gao, Cheng Tan, Tongle Wu, Stan Z. Li
DefectSAM adapts the promptable SAM model to segment defects in infrared pulsed-thermography images. An expert supplies a rough box around a defect, and fine-tuning the image encoder and mask decoder produces a pixel-level outline. In prompt-assisted tests, it reports an IoU of 81.98 on irregular R-type images, versus 65.75 for MedSAM; this measures agreement with labels, not verified defect boundaries.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.