Technology Category
Computer Vision: Motion & TrackingMachine Learning: Time-Series/Data StreamsIntelligent Robots: Multimodal Perception & Sensor Fusion
Interpolation-Based Event Visual Data Filtering Algorithms
🤖 AI Summary
Event-camera data streams suffer from severe noise, and existing methods struggle to balance denoising efficacy with signal fidelity. This paper proposes a lightweight, real-time denoising method that synergistically combines temporal interpolation and infinite impulse response (IIR) filter matrices. We design four novel IIR filtering algorithms, achieving efficient processing with only ~30 KB memory overhead. Our approach jointly leverages spatiotemporal interpolation and adaptive IIR filtering to effectively suppress noise while preserving sharp dynamic event structures without blurring. Evaluated on multiple synthetic and real-world dynamic vision sensor (DVS) datasets, the method achieves an average noise removal rate of 99% while retaining over 95% of valid event signals. To the best of our knowledge, this is the first work to integrate low-order IIR matrix filtering with interpolation mechanisms—delivering high denoising accuracy, minimal computational and memory resources, and strong feasibility for embedded deployment. It establishes a hardware-friendly paradigm for real-time preprocessing of high-resolution event-camera data.
Technology Category
Application Category
📝 Abstract
The field of neuromorphic vision is developing rapidly, and event cameras are finding their way into more and more applications. However, the data stream from these sensors is characterised by significant noise. In this paper, we propose a method for event data that is capable of removing approximately 99% of noise while preserving the majority of the valid signal. We have proposed four algorithms based on the matrix of infinite impulse response (IIR) filters method. We compared them on several event datasets that were further modified by adding artificially generated noise and noise recorded with dynamic vision sensor. The proposed methods use about 30KB of memory for a sensor with a resolution of 1280 x 720 and is therefore well suited for implementation in embedded devices.
Problem
Research questions and friction points this paper is trying to address.
Reducing noise in event camera data streams
Preserving valid signals while filtering noise
Implementing efficient algorithms for embedded devices
Innovation
Methods, ideas, or system contributions that make the work stand out.
Interpolation-based event visual data filtering
Matrix of infinite impulse response filters
Low-memory embedded device implementation
🔎 Similar Papers
No similar papers found.
