DASHA: Decentralized Autofocusing System with Hierarchical Agents

TL;DR

This paper introduces DASHA, a decentralized hierarchical multi-agent reinforcement learning system that autonomously adjusts camera focus to enhance object detection performance under varying conditions.

Contribution

It presents the first no-reference, self-training autofocus system using hierarchical reinforcement learning for improved object detection.

Findings

Significant improvement over traditional detection models.

Effective autonomous camera focus adjustment in diverse conditions.

First method enabling completely no-reference autofocus.

Abstract

State-of-the-art object detection models are frequently trained offline using available datasets, such as ImageNet: large and overly diverse data that are unbalanced and hard to cluster semantically. This kind of training drops the object detection performance should the change in illumination, in the environmental conditions (e.g., rain), or in the lens positioning (out-of-focus blur) occur. We propose a decentralized hierarchical multi-agent deep reinforcement learning approach for intelligently controlling the camera and the lens focusing settings, leading to a significant improvement beyond the capacity of the popular detection models (YOLO, Faster R-CNN, and Retina are considered). The algorithm relies on the latent representation of the camera's stream and, thus, it is the first method to allow a completely no-reference tuning of the camera, where the system trains itself to auto-focus itself.