Reproducibility Aspects of Crack Detection as a Weakly-Supervised Problem: Towards Achieving Less Annotation-Intensive Crack Detectors

摘要

This paper focuses on the reproducibility aspects of our ICPR2020 paper titled Crack Detection as a Weakly-Supervised Problem: Towards Achieving Less Annotation-Intensive Crack Detectors. More specifically, we will describe our efforts in making the proposed framework reproducible, the dataset reproducible, and the experiments reproducible. In addition, we argue that reproducibility is a step toward adopt-able research, which is something all researchers should strive for.