Incremental multi-target domain adaptation for object detection with efficient domain transfer

Abstract

Recent advances in unsupervised domain adaptation have significantly improved the recognition accuracy of CNNs by alleviating the domain shift between (labeled) source and (unlabeled) target data distributions. While the problem of single-target domain adaptation (STDA) for object detection has recently received much attention, multi-target domain adaptation (MTDA) remains largely unexplored, despite its practical relevance in several real-world applications, such as multi-camera video surveillance. Compared to the STDA problem that may involve large domain shifts between complex source and target distributions, MTDA faces additional challenges, most notably the computational requirements and catastrophic forgetting of previously-learned targets, which can depend on the order of target adaptations. STDA for detection can be applied to MTDA by adapting one model per target, or one common model with a mixture of data from target domains. However, these approaches are either costly or inaccurate. The only state-of-art MTDA method specialized for detection learns targets incrementally, one target at a time, and mitigates the loss of knowledge by using a duplicated detection model for knowledge distillation, which is computationally expensive and does not scale well to many domains. In this paper, we introduce an efficient approach for incremental learning that generalizes well to multiple target domains. Our MTDA approach is more suitable for real-world applications since it allows updating the detection model incrementally, without storing data from previous-learned target domains, nor retraining when a new target domain becomes available. Our approach leverages domain discriminators to train a novel Domain Transfer Module (DTM), which only incurs a modest overhead. The DTM transforms source images according to diverse target domains, allowing the model to access a joint representation of previously-learned target domains, and to effectively limit catastrophic forgetting. Our proposed method – called MTDA with DTM (MTDA-DTM) – is compared against state-of-the-art approaches on several MTDA detection benchmarks and Wildtrack, a benchmark for multi-camera pedestrian detection. Results indicate that MTDA-DTM achieves the highest level of detection accuracy across multiple target domains, yet requires significantly fewer computational resources. Our code is available.11https://github.com/Natlem/M-HTCN.