One-Step Event-Driven High-Speed Autofocus

2025-03-03

AI Summary

Conventional autofocus methods suffer from latency and inefficiency in extreme scenarios—such as high-speed motion or low-light conditions—due to their reliance on iterative focus search across multiple frames. Method: This paper proposes the first single-step, event-driven autofocus method, eliminating iterative focus searching entirely. Its core innovation is the Event-based Laplacian Product (ELP), a novel focus metric that reformulates autofocus as a focal-state detection task on a single event stream. We further design an end-to-end, event-camera–specific pipeline (optimized for DAVIS346 and EVK4) that jointly encodes event streams and grayscale Laplacian features. Results: Experiments demonstrate up to 67% reduction in autofocus latency, with focal error reduced by 24× (DAVIS346) and 22× (EVK4). The method significantly enhances real-time performance and focusing accuracy under high-dynamic-range and low-illumination conditions.

Technology Category

• Computer Vision: Motion & Tracking
• Search and Optimization: Learning to Search
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

High-speed autofocus in extreme scenes remains a significant challenge. Traditional methods rely on repeated sampling around the focus position, resulting in ``focus hunting''. Event-driven methods have advanced focusing speed and improved performance in low-light conditions; however, current approaches still require at least one lengthy round of ``focus hunting'', involving the collection of a complete focus stack. We introduce the Event Laplacian Product (ELP) focus detection function, which combines event data with grayscale Laplacian information, redefining focus search as a detection task. This innovation enables the first one-step event-driven autofocus, cutting focusing time by up to two-thirds and reducing focusing error by 24 times on the DAVIS346 dataset and 22 times on the EVK4 dataset. Additionally, we present an autofocus pipeline tailored for event-only cameras, achieving accurate results across a range of challenging motion and lighting conditions. All datasets and code will be made publicly available.

Problem

• High-speed autofocus in extreme scenes
• Eliminate focus hunting in event-driven methods
• Accurate autofocus for event-only cameras

Innovation

• Event Laplacian Product for one-step autofocus
• Combines event data with grayscale Laplacian
• Tailored pipeline for event-only cameras