Constrained Convolutional Neural Networks for Weakly Supervised Segmentation

TL;DR

This paper introduces Constrained CNNs (CCNN), a novel method that incorporates linear constraints derived from image-level tags into CNN training, enabling effective weakly supervised pixel-wise segmentation.

Contribution

The paper proposes a new loss function for CNNs that enforces linear constraints, improving weakly supervised segmentation performance with a flexible and optimizable framework.

Findings

Achieves state-of-the-art results on weakly supervised segmentation

Flexible loss function easily integrated into standard training

Adding more supervision significantly boosts performance

Abstract

We present an approach to learn a dense pixel-wise labeling from image-level tags. Each image-level tag imposes constraints on the output labeling of a Convolutional Neural Network (CNN) classifier. We propose Constrained CNN (CCNN), a method which uses a novel loss function to optimize for any set of linear constraints on the output space (i.e. predicted label distribution) of a CNN. Our loss formulation is easy to optimize and can be incorporated directly into standard stochastic gradient descent optimization. The key idea is to phrase the training objective as a biconvex optimization for linear models, which we then relax to nonlinear deep networks. Extensive experiments demonstrate the generality of our new learning framework. The constrained loss yields state-of-the-art results on weakly supervised semantic image segmentation. We further demonstrate that adding slightly more supervision can greatly improve the performance of the learning algorithm.