Browser Security Notice: For reliable webcam permissions and model downloading, open this page through localhost (e.g. python3 -m http.server 5500) instead of a direct file:// URL.

Live Vision and Detection Dashboard

Real-time PC webcam preview with TensorFlow.js COCO-SSD object detection, FLIR thermal imaging, and visual inspection.

CHECKING LIBRARIES CAMERA OFF RGB STREAM ONLINE OPERATOR VERIFICATION REQUIRED

Video Source Mode:

PC Camera Device:

Primary Visual Monitor (PC Camera)

PC Camera Permission Required

Allow access to your computer webcam for real-time visual monitoring. Audio access is not requested.

The camera stream remains private inside your browser window.

PC CAMERA READY

COCO-SSD Detector Ready

Mirror Preview

URL: Local MediaDevices

FLIR Thermal Radiometric Camera

Heat Signature Detected

Max: 36.8 °C | Min: 22.4 °C

Palette: Iron (Medical) Target: 36.8 °C (Human Heat) Verification: Required

Object Detection Model CHECKING LIBRARIES

TensorFlow.js: Checking Processing Backend: Not selected COCO-SSD: Not loaded

Object Detection Controls

CHECKING LIBRARIES

Min Confidence: 60%

Max Objects:

Processing Quality:

Category Filter:

Bounding Boxes

Object Names

Confidence Scores

Detection List

Primary Recognized Prediction

Lite MobileNet V2

Highest Confidence Object

NO RECOGNISED OBJECT

--% Confidence

Start object detection on an active video stream to analyze live frames.

People: 0

Vehicles: 0

Animals: 0

Equipment: 0

Common Objects: 0

Active Objects Detected (0)

Advisory Predictions

No objects currently detected. Load model and start PC camera to begin inference.

Object names and confidence scores are generated by a general-purpose pretrained COCO-SSD model. Operator verification is required before taking action.