Analysis on DeepLabV3+ Performance for Automatic Steel Defects Detection

TL;DR

This study evaluates DeepLabV3+ with various backbones for automatic steel defect detection, finding ResNet101 and EfficientNet provide the best segmentation performance and training efficiency.

Contribution

It systematically compares different backbones for DeepLabV3+ in steel defect segmentation, highlighting ResNet101 and EfficientNet as optimal choices.

Findings

ResNet101 and EfficientNet achieve IoU scores around 0.57.

DenseNet achieves a lower IoU score of 0.325.

ResNet101 requires the least training time.

Abstract

Our works experimented DeepLabV3+ with different backbones on a large volume of steel images aiming to automatically detect different types of steel defects. Our methods applied random weighted augmentation to balance different defects types in the training set. And then applied DeeplabV3+ model three different backbones, ResNet, DenseNet and EfficientNet, on segmenting defection regions on the steel images. Based on experiments, we found that applying ResNet101 or EfficientNet as backbones could reach the best IoU scores on the test set, which is around 0.57, comparing with 0.325 for using DenseNet. Also, DeepLabV3+ model with ResNet101 as backbone has the fewest training time.