Domain-Specific Suppression for Adaptive Object Detection

TL;DR

This paper introduces a novel domain-specific suppression technique for adaptive object detection, focusing on separating domain-invariant and domain-specific features to improve transferability across different domains.

Contribution

It proposes a new constraint on convolution gradients to detach and suppress domain-specific directions, enhancing domain adaptation in object detection.

Findings

Achieves 10.2-12.2% mAP improvement over state-of-the-art methods

Validates the approach on weather, camera, and synthetic-to-real adaptation tasks

Provides theoretical analysis supporting the effectiveness of the suppression method

Abstract

Domain adaptation methods face performance degradation in object detection, as the complexity of tasks require more about the transferability of the model. We propose a new perspective on how CNN models gain the transferability, viewing the weights of a model as a series of motion patterns. The directions of weights, and the gradients, can be divided into domain-specific and domain-invariant parts, and the goal of domain adaptation is to concentrate on the domain-invariant direction while eliminating the disturbance from domain-specific one. Current UDA object detection methods view the two directions as a whole while optimizing, which will cause domain-invariant direction mismatch even if the output features are perfectly aligned. In this paper, we propose the domain-specific suppression, an exemplary and generalizable constraint to the original convolution gradients in backpropagation to detach the two parts of directions and suppress the domain-specific one. We further validate our theoretical analysis and methods on several domain adaptive object detection tasks, including weather, camera configuration, and synthetic to real-world adaptation. Our experiment results show significant advance over the state-of-the-art methods in the UDA object detection field, performing a promotion of $10.2\sim12.2\%$ mAP on all these domain adaptation scenarios.