This is a series of detections based on opencv and tensorflow running on RaspiberryPi 4B.
When the program is running,it loads the dependencies this program needs such as some Cascade Classifier, a typical object detection model(ssd_mobilenet_v1_coco_2018_01_28) and a typical text detection model(east model). Then it will recognize your gesture(1 to 4), and choose a detection by it.
1->object detection
2->text detection
3->face detection(including eyes and smiles)
4->moving detection
A submodule of this program is /models from Model Garden for TensorFlow and adding a third-party trained model ssd_mobilenet_v1_coco_2018_01_28 ssd_mobilenet_v1_coco_2018_01_28
Dependency | OS Version | RaspiberryPi Version |
---|---|---|
python | 3.7.3 | 3.7.3 |
tensorflow | 2.4.1 | 1.14.0 |
opencv-python | 4.5.1.48 | 4.5.0 |
imutils | 0.5.4 | 0.5.4 |
-
if the tensorflow model didn't work: add the "TensorFlow-ObjectDetection-models.pth" file to python/site-packages.
such as "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages"