An in-depth assessment of convolutional neural networks for rail surface defect detection. (2022). Research, Society and Development, 11(8), e12211830252. https://doi.org/10.33448/rsd-v11i8.30252