Verification of Size Invariance in DNN Activations using Concept Embeddings

TL;DR

This paper improves concept analysis methods for large object detectors, demonstrating that DNN internal representations of body parts are largely size invariant, which has implications for safety-critical applications.

Contribution

It introduces an enhanced Net2Vec approach for sub-object segmentation and applies it to assess size invariance in DNN activations for the first time.

Findings

Body part representations are mostly size invariant.

DNNs show early fusion of information across size categories.

The approach is validated on a new concept dataset.

Abstract

The benefits of deep neural networks (DNNs) have become of interest for safety critical applications like medical ones or automated driving. Here, however, quantitative insights into the DNN inner representations are mandatory. One approach to this is concept analysis, which aims to establish a mapping between the internal representation of a DNN and intuitive semantic concepts. Such can be sub-objects like human body parts that are valuable for validation of pedestrian detection. To our knowledge, concept analysis has not yet been applied to large object detectors, specifically not for sub-parts. Therefore, this work first suggests a substantially improved version of the Net2Vec approach (arXiv:1801.03454) for post-hoc segmentation of sub-objects. Its practical applicability is then demonstrated on a new concept dataset by two exemplary assessments of three standard networks, including the larger Mask R-CNN model (arXiv:1703.06870): (1) the consistency of body part similarity, and (2) the invariance of internal representations of body parts with respect to the size in pixels of the depicted person. The findings show that the representation of body parts is mostly size invariant, which may suggest an early intelligent fusion of information in different size categories.