Convolutional Neural Networks Rarely Learn Shape for Semantic Segmentation

Abstract

Shape learning, or the ability to leverage shape information, could be a desirable property of convolutional neural networks (CNNs) when target objects have specific shapes. While some research on the topic is emerging, there is no systematic study to conclusively determine whether and under what circumstances CNNs learn shape. Here, we present such a study in the context of segmentation networks where shapes are particularly important. We define shape and propose a new behavioral metric to measure the extent to which a CNN utilizes shape information. We then execute a set of experiments with synthetic and real-world data to progressively uncover under which circumstances CNNs learn shape and what can be done to encourage such behavior. We conclude that (i) CNNs do not learn shape in typical settings but rather rely on other features available to identify the objects of interest, (ii) CNNs can learn shape, but only if the shape is the only feature available to identify the object, (iii) sufficiently large receptive field size relative to the size of target objects is necessary for shape learning; (iv) a limited set of augmentations can encourage shape learning; (v) learning shape is indeed useful in the presence of out-of-distribution data.

arXiv commentAccepted by Pattern Recognition, reformatted for Arxiv for readability

References in corpus (7)

• Ensemble Adversarial Training: Attacks and Defenses
• More Diverse Means Better: Multimodal Deep Learning Meets Remote Sensing Imagery Classification
• An Augmented Linear Mixing Model to Address Spectral Variability for Hyperspectral Unmixing
• Association of genomic subtypes of lower-grade gliomas with shape features automatically extracted by a deep learning algorithm
• The Origins and Prevalence of Texture Bias in Convolutional Neural Networks
• Detection of masses and architectural distortions in digital breast tomosynthesis: a publicly available dataset of 5,060 patients and a deep learning model
• What shapes feature representations? Exploring datasets, architectures, and training