Real-time Human Detection in Computer Vision — Part 2

Modern approaches for Human Detection

Human Detection using “Faster RCN Inception V2 COCO” model available with Tensorflow Object Detection API. Test Video from “Coarse Gaze Estimation in Visual Surveillance Project” by University of Oxford

They are “Deep Convolution Neural Networks”

They are “Multi-class Object Detectors”

Human Detection using Tensorflow Object Detection API

Setting up a Basic Human Detector

Human Detection using faster_rcnn_inception_v2_coco Model (Threshold = 0.7)

Test Bench

Comparison among models from Tensorflow Detection Model Zoo

SSD Mobilenet V1 COCO Model

Human Detection using ssd_mobilenet_v1_coco model (Left: Threshold = 0.2; Center: Threshold = 0.3; Right: Threshold = 0.5)

Faster RCN Inception V2 COCO Model

Faster RCN NAS COCO Model

Human Detection using Faster RCN NAS COCO model (threshold = 0.5)

Improvements from earlier approaches

Haar cascade (left), SSD Mobilenet V1 (center) and Faster RCN Inception V2 (right)

Drawbacks compared to earlier approaches

Conclusion

Reference

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade