DNN - Object Detection
Object detection will recognize various classes within an image and provide a bounding box, class label and confidence value.
See Deep Learning in OpenCV for information about supported frameworks and supported layers.
Some tested models with parameters:
| Model | Config | Classes | Size (W x H) | Mean | Scale | RGB |
|---|---|---|---|---|---|---|
| yolo v3 model | yolo v3 config | classes yolov3 | 416 x 416 | 0, 0, 0 | 0.00392 | true |
| tiny yolo voc model | tiny yolo voc config | classes pascal voc | 416 x416 | 0, 0, 0 | 0.00392 | true |
| ssd caffe model | mobilenet ssd caffe config | classes pascal voc | 300 x 300 | 127.5, 127.5, 127.5 | 0.007843 | false |
| ssd tensorflow model | mobilenet ssd tensorflow config | classes coco | 300 x 300 | 0, 0, 0 | 1.0 | true |
| faster rcnn model | faster rcnn config | classes coco | 800 x 600 | 0, 0, 0 | 1.0 | true |
| OpenCV Face Detector | face detector config | No Class File | 300 x 300 | 104, 177, 123 | 1.0 | false |