TinyGLASS enables real-time in-sensor anomaly detection on edge devices

Researchers have developed TinyGLASS, a lightweight adaptation of the GLASS framework for real-time, self-supervised anomaly detection on resource-constrained edge devices. This new architecture utilizes a compact ResNet-18 backbone and incorporates modifications for static graph tracing and INT8 quantization, enabling deployment on in-sensor processors like the Sony IMX500. TinyGLASS achieves significant parameter compression while maintaining competitive performance, operating at 20 FPS within strict memory limits and demonstrating low power consumption. AI

IMPACT Enables real-time anomaly detection on edge devices, potentially improving industrial quality control and efficiency.

RANK_REASON The cluster describes a new research paper detailing a novel model adaptation for in-sensor anomaly detection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Luigi Capogrosso Ph.D.
• MMS Dataset
• MVTec-AD
• ResNet-18
• Sony IMX500
• Sony Model Compression Toolkit
• TinyGLASS
• WideResNet-50

• model release
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →