Python based UI in PyQt5 for choosing images and using ML algorithms (FRCNN, SSD Inception, SSD MobileNet) to label them. Various options are also provided.
Sample ScreenRecording- https://drive.google.com/open?id=1NnT5_huRoxYR65tYrf49l9cvkUHmY-SQ
To use this project, just
- Download and load models from the links and copy in actual_model folder Links - i) http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_v2_coco_2018_01_28.tar.gz ii) http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2018_01_28.tar.gz iii) http://download.tensorflow.org/models/object_detection/ssd_inception_v2_coco_2018_01_28.tar.gz Example directory structure is /actual_model/faster_rcnn_inception_v2_coco_2018_01_28/ containing all the files
- run Recognize.py (in this implementation, model gets load after clicking Detect) or Recognize2.py (here, models are preloaded, high on RAM but faster) using python3.