MILA: Memory-Based Instance-Level Adaptation for Cross-Domain Object Detection

20 Nov 2023  ·  Onkar Krishna; Hiroki Ohashi; Saptarshi Sinha ·

Cross-domain object detection is challenging, and it involves aligning labeled source and unlabeled target domains. Previous approaches have used adversarial training to align features at both image-level and instance-level. At the instance level, finding a suitable source sample that aligns with a target sample is crucial. A source sample is considered suitable if it differs from the target sample only in domain, without differences in unimportant characteristics such as orientation and color, which can hinder the model’s focus on aligning the domain difference. However, existing instance-level feature alignment methods struggle to find suitable source instances because their search scope is limited to mini-batches. Mini-batches are often so small in size that they do not always contain suitable source instances. The insufficient diversity of mini-batches becomes problematic particularly when the target instances have high intra-class variance. To address this issue, we propose a memory-based instance-level domain adaptation framework. Our method aligns a target instance with the most similar source instance of the same category retrieved from a memory storage. Specifically, we introduce a memory module that dynamically stores the pooled features of all labeled source instances, categorized by their labels. Additionally, we introduce a simple yet effective memory retrieval module that retrieves a set of matching memory slots for target instances. Our experiments on various domain shift scenarios demonstrate that our approach outperforms existing non-memory-based methods significantly.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Domain Adaptation Comic2k MILA mAP 44.6 # 1
Domain Adaptation Foggy Cityscapes MILA mAP 50.6 # 1
Object Detection Pascal VOC to Clipart1K MILA mAP 49.9 # 1
Domain Adaptation Sim10k MILA mAP 57.4 # 1