Region-based semantic segmentation with end-to-end training

26 Jul 2016  ·  Holger Caesar, Jasper Uijlings, Vittorio Ferrari ·

We propose a novel method for semantic segmentation, the task of labeling each pixel in an image with a semantic class. Our method combines the advantages of the two main competing paradigms. Methods based on region classification offer proper spatial support for appearance measurements, but typically operate in two separate stages, none of which targets pixel labeling performance at the end of the pipeline. More recent fully convolutional methods are capable of end-to-end training for the final pixel labeling, but resort to fixed patches as spatial support. We show how to modify modern region-based approaches to enable end-to-end training for semantic segmentation. This is achieved via a differentiable region-to-pixel layer and a differentiable free-form Region-of-Interest pooling layer. Our method improves the state-of-the-art in terms of class-average accuracy with 64.0% on SIFT Flow and 49.9% on PASCAL Context, and is particularly accurate at object boundaries.

PDF Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Semantic Segmentation PASCAL Context RBE2E mIoU 32.5 # 62
Mean Accuracy 49.9 # 1
Pixel Accuracy 62.4 # 1
Semantic Segmentation SIFT-flow RBE2E Mean Accuracy 64 # 1
Pixel Accuracy 84.3 # 1

Methods


No methods listed for this paper. Add relevant methods here