Method Details


Details for method 'SSMA'

 

Method overview

name SSMA
challenge pixel-level semantic labeling
details Learning to reliably perceive and understand the scene is an integral enabler for robots to operate in the real-world. This problem is inherently challenging due to the multitude of object types as well as appearance changes caused by varying illumination and weather conditions. Leveraging complementary modalities can enable learning of semantically richer representations that are resilient to such perturbations. Despite the tremendous progress in recent years, most multimodal convolutional neural network approaches directly concatenate feature maps from individual modality streams rendering the model incapable of focusing only on the relevant complementary information for fusion. To address this limitation, we propose a mutimodal semantic segmentation framework that dynamically adapts the fusion of modality-specific features while being sensitive to the object category, spatial location and scene context in a self-supervised manner. Specifically, we propose an architecture consisting of two modality-specific encoder streams that fuse intermediate encoder representations into a single decoder using our proposed SSMA fusion mechanism which optimally combines complementary features. As intermediate representations are not aligned across modalities, we introduce an attention scheme for better correlation. Extensive experimental evaluations on the challenging Cityscapes, Synthia, SUN RGB-D, ScanNet and Freiburg Forest datasets demonstrate that our architecture achieves state-of-the-art performance in addition to providing exceptional robustness in adverse perceptual conditions. Please refer to https://arxiv.org/abs/1808.03833 for details. A live demo on various datasets can be viewed at http://deepscene.cs.uni-freiburg.de
publication Self-Supervised Model Adaptation for Multimodal Semantic Segmentation
Abhinav Valada, Rohit Mohan, Wolfram Burgard
arXiv
https://arxiv.org/abs/1808.03833
project page / code http://deepscene.cs.uni-freiburg.de
used Cityscapes data fine annotations, coarse annotations, stereo
used external data ImageNet
runtime n/a
subsampling no
submission date January, 2019
previous submissions

 

Average results

Metric Value
IoU Classes 82.312
iIoU Classes 62.2501
IoU Categories 91.5078
iIoU Categories 81.7139

 

Class results

Class IoU iIoU
road 98.6664 -
sidewalk 86.884 -
building 93.605 -
wall 57.8519 -
fence 63.4302 -
pole 68.938 -
traffic light 77.1464 -
traffic sign 81.1373 -
vegetation 93.8571 -
terrain 73.0615 -
sky 95.3172 -
person 87.4316 72.6122
rider 73.7845 52.3686
car 96.3584 91.4028
truck 81.1375 47.834
bus 93.4868 58.08
train 89.9538 58.6083
motorcycle 73.5405 51.6583
bicycle 78.3401 65.437

 

Category results

Category IoU iIoU
flat 98.7282 -
nature 93.4833 -
object 75.2108 -
sky 95.3172 -
construction 93.9055 -
human 87.8295 73.6083
vehicle 96.0798 89.8195

 

Links

Download results as .csv file

Benchmark page