Technical news and analysis for intelligent video surveillance
NVIDIA AI-RAN transforms telecom infrastructure into a distributed edge compute layer for continuous inference on real-time video/sensor streams via 5G to edge ...
Flasheye proposes a hybrid architecture using LiDAR for 3D spatial awareness and event detection to intelligently control PTZ cameras, directing precise pan-til...
Compares large vision models like GPT-4o, YOLOv8n, and DeepMind Flamingo for surveillance applications including activity recognition, anomaly detection, and co...
TII's Falcon-H1-Tiny SLMs (90M params, hybrid architecture) deliver edge reasoning, multilingual capabilities, and instruction-following rivaling larger models ...
![Visual Anomaly Detection with Anomalib: A Hands-On Guide [2026] | Datature Blog](https://cdn.prod.website-files.com/67d7ebcca3cabb623904dcde/69b9dbeb106d1f2d400ad362_Datature%20Blog%20Thumbnails%20(13).png)
Hands-on guide to visual anomaly detection using Anomalib's 23 algorithms, focusing on one-class classification with PatchCore, PaDiM, and EfficientAd; trains s...
Details COTS edge AI accelerators like EdgeCortix SAKURA-II for SWaP-C constrained defense/space ISR, supporting YOLOv8/11 for real-time object detection/tracki...
Spotlights AI camera developers emphasizing TEEs, ISP-level PII anonymization, and cybersecurity audits for gov/healthcare, with SQUAD's full-stack edge AI enab...
Field tests highlight edge AI cameras using NPUs like STM32N6 for sub-300ms object detection (human/vehicle/pet) to suppress false alarms from PIR triggers, shi...
Guides edge deployment of CV models with focus on real-time pipelines, telemetry for inference latency/memory monitoring, and lightweight anomaly detection in r...
FPT's AI surveillance for a Taiwanese manufacturer uses NVIDIA RTX 4080 GPUs and NVILA Vision Language Model for real-time multi-camera behavior tracking and ha...
Report analyzes smart building trends including on-camera edge AI and vision-language models for video search in physical security, alongside IT-OT convergence ...
Discusses edge AI for robotics emphasizing sensor fusion (vision + force/motion/depth) for robust perception, anomaly detection via normal behavior modeling to ...
Details edge AI deployments in healthcare (HIPAA-compliant local inference on cameras to avoid raw data transmission and ensure reliability during network outag...
Details 2026 anomaly detection architectures learning normal scene distributions (motion vectors, dwell time, trajectories) to flag real-time anomalies like loi...
March Networks' AI Smart Search enables natural language queries for rapid video event retrieval, accelerating investigations and compliance checks. Their Long ...
Edge AI enables real-time object detection/classification at 30+ fps, autonomous obstacle avoidance under 100ms, and anomaly detection via baseline learning on ...
ForaSoft's V.A.L.T platform enables brownfield integration of legacy pre-ONVIF/RTSP cameras via pluggable analytics layer, supporting HD streaming from 2500+ ca...
Explores Edge AI deployments in surveillance cameras for local inference on object detection, occupancy, and anomaly tasks, transmitting only structured metadat...
SurveilSmartAI is a unified edge-deployable architecture integrating real-time object detection, human tracking, and anomaly recognition, addressing scale-aware...
Outlines a resilient edge AI stack for factory floors: camera/sensor input, preprocessing, local inference, decision logic, buffering, and async cloud sync for ...