Method Details

Details for method 'Adaptive Affinity Field on PSPNet'

name	Adaptive Affinity Field on PSPNet
challenge	pixel-level semantic labeling
details	Existing semantic segmentation methods mostly rely on per-pixel supervision, unable to capture structural regularity present in natural images. Instead of learning to enforce semantic labels on individual pixels, we propose to enforce affinity field patterns in individual pixel neighbourhoods, i.e., the semantic label patterns of whether neighbouring pixels are in the same segment should match between the prediction and the ground-truth. The affinity fields characterize geometric relationships within the image, such as "motorcycles have round wheels". We further develop a novel method for learning the optimal neighbourhood size for each semantic category, with an adversarial loss that optimizes over worst-case scenarios. Unlike the common Conditional Random Field (CRF) approaches, our adaptive affinity field (AAF) method has no extra parameters during inference, and is less sensitive to appearance changes in the image.
publication	Adaptive Affinity Field for Semantic Segmentation Tsung-Wei Ke, Jyh-Jing Hwang, Ziwei Liu, Stella X. Yu ECCV 2018 https://arxiv.org/abs/1803.10335
project page / code	https://github.com/twke18/Adaptive_Affinity_Fields
used Cityscapes data	fine annotations
used external data	ImageNet
runtime	n/a
subsampling	no
submission date	May, 2018
previous submissions