Attentional Network for Visual Object Detection

TL;DR

This paper introduces an attention-based deep neural network for visual object detection, inspired by human vision, which adaptively focuses on different image regions to improve detection accuracy.

Contribution

The paper presents a novel attention mechanism integrated into deep networks for object detection, trained with reinforcement learning, outperforming baseline models.

Findings

Outperforms baseline detection models on standard benchmarks.

Uses reinforcement learning due to lack of explicit attention ground truth.

Imitates human visual attention to enhance detection performance.

Abstract

We propose augmenting deep neural networks with an attention mechanism for the visual object detection task. As perceiving a scene, humans have the capability of multiple fixation points, each attended to scene content at different locations and scales. However, such a mechanism is missing in the current state-of-the-art visual object detection methods. Inspired by the human vision system, we propose a novel deep network architecture that imitates this attention mechanism. As detecting objects in an image, the network adaptively places a sequence of glimpses of different shapes at different locations in the image. Evidences of the presence of an object and its location are extracted from these glimpses, which are then fused for estimating the object class and bounding box coordinates. Due to lacks of ground truth annotations of the visual attention mechanism, we train our network using a reinforcement learning algorithm with policy gradients. Experiment results on standard object detection benchmarks show that the proposed network consistently outperforms the baseline networks that does not model the attention mechanism.