On the Road to Online Adaptation for Semantic Image Segmentation
Authors:
- 1 June 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 19162-19173
Abstract
We propose a new problem formulation and a corresponding evaluation framework to advance research on unsupervised domain adaptation for semantic image segmentation. The overall goal is fostering the development of adaptive learning systems that will continuously learn, without supervision, in ever-changing environments. Typical protocols that study adaptation algorithms for segmentation models are limited to few domains, adaptation happens offline, and human intervention is generally required, at least to annotate data for hyperparameter tuning. We argue that such constraints are incompatible with algorithms that can continuously adapt to different real-world situations. To address this, we propose a protocol where models need to learn online, from sequences of temporally correlated images, requiring continuous, frame-by-frame adaptation. We accompany this new protocol with a variety of baselines to tackle the proposed formulation, as well as an extensive analysis of their behaviors, which can serve as a starting point for future research.Keywords
All Related Versions
This publication has 56 references indexed in Scilit:
- Image Style Transfer Using Convolutional Neural Networks2016
- The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes2016
- The Cityscapes Dataset for Semantic Urban Scene Understanding2016
- Conditional Random Fields as Recurrent Neural Networks2015
- Learning Deconvolution Network for Semantic Segmentation2015
- Scene labeling with LSTM recurrent neural networks2015
- Fully convolutional networks for semantic segmentation2015
- Online Domain Adaptation for Multi-Object Tracking2015
- Incremental Domain Adaptation of Deformable Part-based Models2014
- Learning representations by back-propagating errorsNature, 1986