Multi-Domain Incremental Learning for Semantic Segmentation

• Published in: 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) pp. 2080 - 2090
• Main Authors: Garg, Prachi, Saluja, Rohit, Balasubramanian, Vineeth N, Arora, Chetan, Subramanian, Anbumani, Jawahar, C.V.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-01-2022
• Subjects: Computer architecture; Computer vision; Deep Learning Transfer; Few-shot; Semi- and Un- supervised Learning; Interference; Roads; Semantics; Stability analysis; Visualization
• ISSN: 2642-9381
• DOI: 10.1109/WACV51458.2022.00214

Recent efforts in multi-domain learning for semantic segmentation attempt to learn multiple geographical datasets in a universal, joint model. A simple fine-tuning experiment performed sequentially on three popular road scene segmentation datasets demonstrates that existing segmentation frameworks fail at incrementally learning on a series of visually disparate geographical domains. When learning a new domain, the model catastrophically forgets previously learned knowledge. In this work, we pose the problem of multi-domain incremental learning for semantic segmentation. Given a model trained on a particular geographical domain, the goal is to (i) incrementally learn a new geographical domain, (ii) while retaining performance on the old domain, (iii) given that the previous domain's dataset is not accessible. We propose a dynamic architecture that assigns universally shared, domain-invariant parameters to capture homogeneous semantic features present in all domains, while dedicated domain-specific parameters learn the statistics of each domain. Our novel optimization strategy helps achieve a good balance between retention of old knowledge (stability) and acquiring new knowledge (plasticity). We demonstrate the effectiveness of our proposed solution on domain incremental settings pertaining to real-world driving scenes from roads of Germany (Cityscapes), the United States (BDD100k), and India (IDD). 1