Performance Evaluation of a Raspberry Pi running Deep Learning Mobilenet with a shrinked resolution of 640×480 raspicam for detecting 10 pre-trained categories/classes: bird, cow, cat, dog, sheep, horse, person, pottedplant, tvmonitor, chair.

The results show multi-objects classification and localisation in a complex scene within 80ms on a Single Board Computer Raspberry Pi.

Hardware cost: <150 USD.

Time cost: A lot of fun and perseverance.