Cookies

We use cookies to ensure that we give you the best experience on our website. By continuing to browse this repository, you give consent for essential cookies to be used. You can read more about our Privacy and Cookie Policy.


Durham Research Online
You are in:

“Just Drive”: Colour Bias Mitigation for Semantic Segmentation in the Context of Urban Driving

Stelling, Jack and Atapour-Abarghouei, Amir (2021) '“Just Drive”: Colour Bias Mitigation for Semantic Segmentation in the Context of Urban Driving.', 2021 IEEE International Conference on Big Data (IEEE BigData 2021).

Abstract

Biases can filter into AI technology without our knowledge. Oftentimes, seminal deep learning networks champion increased accuracy above all else. In this paper, we attempt to alleviate biases encountered by semantic segmentation models in urban driving scenes, via an iteratively trained unlearning algorithm. Convolutional neural networks have been shown to rely on colour and texture rather than geometry. This raises issues when safety-critical applications, such as self-driving cars, encounter images with covariate shift at test time - induced by variations such as lighting changes or seasonality. Conceptual proof of bias unlearning has been shown on simple datasets such as MNIST. However, the strategy has never been applied to the safety-critical domain of pixel-wise semantic segmentation of highly variable training data - such as urban scenes. Trained models for both the baseline and bias unlearning scheme have been tested for performance on colour-manipulated validation sets showing a disparity of up to 85.50% in mIoU from the original RGB images - confirming segmentation networks strongly depend on the colour information in the training data to make their classification. The bias unlearning scheme shows improvements of handling this covariate shift of up to 61% in the best observed case - and performs consistently better at classifying the “human” and “vehicle” classes compared to the baseline model

Item Type:Conference item (Paper)
Full text:(AM) Accepted Manuscript
Download PDF
(21907Kb)
Status:Peer-reviewed
Publisher Web site:https://bigdataieee.org/BigData2021/index.html
Publisher statement:© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Date accepted:11 November 2021
Date deposited:03 December 2021
Date of first online publication:2021
Date first made open access:03 December 2021

Save or Share this output

Export:
Export
Look up in GoogleScholar