Towards Universal Object Detection by Domain Attention

TL;DR

This paper introduces a universal object detection system capable of handling diverse image domains without prior domain knowledge, using novel adaptation layers and domain-attention mechanisms, outperforming existing methods.

Contribution

The paper presents a new universal object detection model with domain-attention and adaptation layers that works across multiple domains without prior domain information.

Findings

Outperforms individual and multi-domain detectors on 11 datasets

Uses only 1.3x more parameters than a single-domain baseline

Achieves state-of-the-art results on a new universal detection benchmark

Abstract

Despite increasing efforts on universal representations for visual recognition, few have addressed object detection. In this paper, we develop an effective and efficient universal object detection system that is capable of working on various image domains, from human faces and traffic signs to medical CT images. Unlike multi-domain models, this universal model does not require prior knowledge of the domain of interest. This is achieved by the introduction of a new family of adaptation layers, based on the principles of squeeze and excitation, and a new domain-attention mechanism. In the proposed universal detector, all parameters and computations are shared across domains, and a single network processes all domains all the time. Experiments, on a newly established universal object detection benchmark of 11 diverse datasets, show that the proposed detector outperforms a bank of individual detectors, a multi-domain detector, and a baseline universal detector, with a 1.3x parameter increase over a single-domain baseline detector. The code and benchmark will be released at http://www.svcl.ucsd.edu/projects/universal-detection/.