Assessing The Importance Of Colours For CNNs In Object Recognition
Aditya Singh, Alessandro Bay, Andrea Mirabile

TL;DR
This study empirically explores the role of colours in CNN-based object recognition, revealing that CNNs often depend heavily on colour cues, with dependency varying across datasets and training methods.
Contribution
It provides the first comprehensive analysis of colour importance in CNNs for object recognition, highlighting how training from scratch increases colour reliance and pre-training reduces it.
Findings
CNNs rely heavily on colour information for predictions.
Dependency on colour varies across datasets.
Pre-training reduces colour dependence.
Abstract
Humans rely heavily on shapes as a primary cue for object recognition. As secondary cues, colours and textures are also beneficial in this regard. Convolutional neural networks (CNNs), an imitation of biological neural networks, have been shown to exhibit conflicting properties. Some studies indicate that CNNs are biased towards textures whereas, another set of studies suggests shape bias for a classification task. However, they do not discuss the role of colours, implying its possible humble role in the task of object recognition. In this paper, we empirically investigate the importance of colours in object recognition for CNNs. We are able to demonstrate that CNNs often rely heavily on colour information while making a prediction. Our results show that the degree of dependency on colours tend to vary from one dataset to another. Moreover, networks tend to rely more on colours if…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsVisual Attention and Saliency Detection · Neural Networks and Applications · Visual perception and processing mechanisms
