Automated Annotation of Shearographic Measurements Enabling Weakly Supervised Defect Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Plassmann, Jessica, Schuler, Nicolas, Schuth, Michael, von Freymann, Georg
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910160048357376
author Plassmann, Jessica
Schuler, Nicolas
Schuth, Michael
von Freymann, Georg
author_facet Plassmann, Jessica
Schuler, Nicolas
Schuth, Michael
von Freymann, Georg
contents Shearography is an interferometric technique sensitive to surface displacement gradients, providing high sensitivity for detecting subsurface defects in safety-critical components. A key limitation to industrial adoption is the lack of high-quality annotated datasets, since manual labeling remains labor-intensive, subjective, and difficult to standardize. We present an automated labeling pipeline that generates candidate defect bounding boxes with Grounded DINO, refines them using SAM masks, and exports YOLO-format labels for downstream detector training. Quantitative evaluation shows the generated boxes are suitable for weakly supervised learning, while high-resolution masks provide qualitative visualization. This approach reduces manual effort and supports scalable dataset creation for robust industrial defect detection.
format Preprint
id arxiv_https___arxiv_org_abs_2512_06171
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Automated Annotation of Shearographic Measurements Enabling Weakly Supervised Defect Detection
Plassmann, Jessica
Schuler, Nicolas
Schuth, Michael
von Freymann, Georg
Computer Vision and Pattern Recognition
Shearography is an interferometric technique sensitive to surface displacement gradients, providing high sensitivity for detecting subsurface defects in safety-critical components. A key limitation to industrial adoption is the lack of high-quality annotated datasets, since manual labeling remains labor-intensive, subjective, and difficult to standardize. We present an automated labeling pipeline that generates candidate defect bounding boxes with Grounded DINO, refines them using SAM masks, and exports YOLO-format labels for downstream detector training. Quantitative evaluation shows the generated boxes are suitable for weakly supervised learning, while high-resolution masks provide qualitative visualization. This approach reduces manual effort and supports scalable dataset creation for robust industrial defect detection.
title Automated Annotation of Shearographic Measurements Enabling Weakly Supervised Defect Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.06171