Model-agnostic Body Part Relevance Assessment for Pedestrian Detection
This work addresses the problem of inefficient explainability analyses for computer vision models, particularly in pedestrian detection, by proposing a more robust and efficient method, though it appears incremental as it builds upon existing sampling-based approaches.
The paper tackles the inefficiency of sampling-based explanation methods like KernelSHAP for complex models with large inputs, such as in pedestrian detection, by introducing a framework for body part relevance assessment and a novel sampling-based method that improves robustness at lower sampling sizes, making it more efficient for large-scale datasets.
Model-agnostic explanation methods for deep learning models are flexible regarding usability and availability. However, due to the fact that they can only manipulate input to see changes in output, they suffer from weak performance when used with complex model architectures. For models with large inputs as, for instance, in object detection, sampling-based methods like KernelSHAP are inefficient due to many computation-heavy forward passes through the model. In this work, we present a framework for using sampling-based explanation models in a computer vision context by body part relevance assessment for pedestrian detection. Furthermore, we introduce a novel sampling-based method similar to KernelSHAP that shows more robustness for lower sampling sizes and, thus, is more efficient for explainability analyses on large-scale datasets.