Topological Regularization for Dense Prediction

Abstract

Dense prediction tasks such as depth perception and semantic segmentation are important applications in computer vision that have a concrete topological description in terms of partitioning an image into connected components or estimating a function with a small number of local extrema corresponding to objects in the image. We develop a form of topological regularization based on persistent homology that can be used in dense prediction tasks with these topological descriptions. Experimental results show that the output topology can also appear in the internal activations of trained neural networks which allows for a novel use of topological regularization to the internal states of neural networks during training, reducing the computational cost of the regularization. We demonstrate that this topological regularization of internal activations leads to improved convergence and test benchmarks on several problems and architectures.

References in corpus (10)

• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
• From Big to Small: Multi-Scale Local Planar Guidance for Monocular Depth Estimation
• High Quality Monocular Depth Estimation via Transfer Learning
• Topology-Preserving Deep Image Segmentation
• A Topological Regularizer for Classifiers via Persistent Homology
• DIODE: A Dense Indoor and Outdoor DEpth Dataset
• Self-Supervised Visual Representation Learning from Hierarchical Grouping
• Exploring Stereotypes and Biased Data with the Crowd
• Understanding and Evaluating Racial Biases in Image Captioning