A object detection application for android devices to detect pedestrian traffic ligth phases
| mAP | [email protected] | AR@100 |
|---|---|---|
| 44.31 % | 84.99 % | 53.73 % |
A on TensorFlowLite OD demo app based implementation of a object detection model for pedestrian traffic light phases. Modified to inform the user about the currently stable detection of a light phase via auditive and tactile output. The app reached 11.82 FPS on a OnePlus6.
| speed | size |
|---|---|
| 11.82 FPS | ~30 MB |