In this work we tackle the problem of unsupervised domain adaptation for semantic segmentation. One pipeline is to sequentially train image-Translation model and the final task segmentation model. In such pipeline, image translation is aimed to generate the translated source-domain images which are visually similar to the target-domain images and then the final task model is trained using the translated images and its corresponding groundtruth. However, the visually optimal translated-images are not necessarily optimal for the final task of segmenting the target-domain images. Thus we propose a Task-in-All pipeline for unsupervised domain adaptation on semantic segmentation, which incorporates image translation and final segmentation task into an end-To-end training pipeline. Our aim is to generate the translated images which better assists the final task, instead of just being visually similar to the target domain images. We show that in the task of adapting from GTA5 to Cityscapes dataset, the segmentation performance of our Task-in-All pipeline outperforms the sequentially training pipeline, with simpler model structure and less training complexity.
|Title of host publication||2019 IEEE International Conference on Visual Communications and Image Processing (VCIP 2019)|
|Editors||Mark Pickering, Qiang Wu, Lei Wang, Jiaying Liu|
|Place of Publication||Piscataway NJ USA|
|Publisher||IEEE, Institute of Electrical and Electronics Engineers|
|Number of pages||4|
|Publication status||Published - 2019|
|Event||IEEE Visual Communications and Image Processing 2019 - Sydney, Australia|
Duration: 1 Dec 2019 → 4 Dec 2019
Conference number: 34th
|Conference||IEEE Visual Communications and Image Processing 2019|
|Abbreviated title||VCIP 2019|
|Period||1/12/19 → 4/12/19|
- Image translation
- semantic segmentation.