English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The accuracy and fairness of perception systems in autonomous driving arecrucial, particularly for vulnerable road users. Mainstream research has lookedinto improving the performance metrics for classification accuracy. However,the hidden traits of bias inheritance in the AI models, class imbalances anddisparities in the datasets are often overlooked. In this context, our studyexamines the class imbalances for vulnerable road users by focusing on classdistribution analysis, performance evaluation, and bias impact assessment. Weidentify the concern of imbalances in class representation, leading topotential biases in detection accuracy. Utilizing popular CNN models and VisionTransformers (ViTs) with the nuScenes dataset, our performance evaluationreveals detection disparities for underrepresented classes. We propose amethodology for model optimization and bias mitigation, which includes dataaugmentation, resampling, and metric-specific learning. Using the proposedmitigation approaches, we see improvement in IoU(%) and NDS(%) metrics from71.3 to 75.6 and 80.6 to 83.7 respectively, for the CNN model. Similarly, forViT, we observe improvement in IoU and NDS metrics from 74.9 to 79.2 and 83.8to 87.1 respectively. This research contributes to developing more reliablemodels and datasets, enhancing inclusiveness for minority classes.

Analyzing and Mitigating Bias for Vulnerable Classes: Towards Balanced Representation in Dataset