Abstract:Semantic segmentation (SS) aims to classify each pixel into one of the pre-defined classes. This task plays an important role in self-driving cars and autonomous drones. In SS, many works have shown that most misclassified pixels are commonly near object boundaries with high uncertainties. However, existing SS loss functions are not tailored to handle these uncertain pixels during training, as these pixels are usually treated equally as confidently classified pixels and cannot be embedded with arbitrary low distortion in Euclidean space, thereby degenerating the performance of SS. To overcome this problem, this paper designs a Hyperbolic Uncertainty Loss (HyperUL), which dynamically highlights the misclassified and high-uncertainty pixels in Hyperbolic space during training via the hyperbolic distances. The proposed HyperUL is model agnostic and can be easily applied to various neural architectures. After employing HyperUL to three recent SS models, the experimental results on Cityscapes, UAVid, and ACDC datasets reveal that the segmentation performance of existing SS models can be consistently improved. Additionally, reliable measurement of model uncertainty plays a key role in real-world applications such as autonomous controls of vehicles and drones. To meet this requirement, we propose the Hyperbolic Uncertainty Estimation method, which is easily implemented by only post-processing the generated Hyperbolic embeddings. By this approach, we can calculate the uncertainty values almost for free. Quantitative and qualitative results on Cityscapes, UAVid, and ACDC datasets verify that our proposed uncertainty estimation method usually outputs more meaningful results compared with popular MC-dropout and ensembling methods.

Hyperbolic Image Segmentation

Piecewise Flat Embedding for Image Segmentation

Seg-HGNN: Unsupervised and Light-Weight Image Segmentation with Hyperbolic Graph Neural Networks

Hyperbolic Deep Learning in Computer Vision: A Survey

Hyperbolic Geometry in Computer Vision: A Survey

HyperSpace: Hypernetworks for spacing-adaptive image segmentation

Hyperbolic Distance Matrices

Factor Space and Spectrum for Medical Hyperspectral Image Segmentation

Unsupervised Multi-Class Co-Segmentation via Joint-Cut Over $L_{1}$ -Manifold Hyper-Graph of Discriminative Image Regions

Intrinsic Image Segmentation on a Riemannian Manifold

Geodesic Paths for Image Segmentation With Implicit Region-Based Homogeneity Enhancement

Kernel Methods in Hyperbolic Spaces

Hyperbolic Representation Learning: Revisiting and Advancing

Hyperbolic Uncertainty Aware Semantic Segmentation

Hypernet-Ensemble Learning of Segmentation Probability for Medical Image Segmentation with Ambiguous Labels

Flattening the Parent Bias: Hierarchical Semantic Segmentation in the Poincaré Ball

Image Segmentation Based On Fuzzy Hypergraph Model

A Generalized Geodesic Voting Framework for Interactive Image Segmentation

Fully Hyperbolic Convolutional Neural Networks for Computer Vision

On Hyperbolic Embeddings in 2D Object Detection