Unintended Bias in 2D+ Image Segmentation and Its Effect on Attention Asymmetry
This addresses reliability issues in specialized domains like biomedical imaging, though it is incremental as it builds on known transfer learning problems.
The study investigated how pretrained models introduce unintended biases in image segmentation, particularly for biomedical imaging, causing attention asymmetry and compromising performance. The proposed methods to neutralize these biases showed promising results for improving model explainability while maintaining pretrained benefits.
Supervised pretrained models have become widely used in deep learning, especially for image segmentation tasks. However, when applied to specialized datasets such as biomedical imaging, pretrained weights often introduce unintended biases. These biases cause models to assign different levels of importance to different slices, leading to inconsistencies in feature utilization, which can be observed as asymmetries in saliency map distributions. This transfer of color distributions from natural images to non-natural datasets can compromise model performance and reduce the reliability of results. In this study, we investigate the effects of these biases and propose strategies to mitigate them. Through a series of experiments, we test both pretrained and randomly initialized models, comparing their performance and saliency map distributions. Our proposed methods, which aim to neutralize the bias introduced by pretrained color channel weights, demonstrate promising results, offering a practical approach to improving model explainability while maintaining the benefits of pretrained models. This publication presents our findings, providing insights into addressing pretrained weight biases across various deep learning tasks.