Let's now see the main components of using pre-trained TensorFlow object detection models for inference in the Python notebook. First, some key constants are defined:
MODEL_NAME = 'ssd_mobilenet_v1_coco_2017_11_17'MODEL_FILE = MODEL_NAME + '.tar.gz'DOWNLOAD_BASE = 'http://download.tensorflow.org/models/object_detection/'PATH_TO_CKPT = MODEL_NAME + '/frozen_inference_graph.pb'PATH_TO_LABELS = os.path.join('data', 'mscoco_label_map.pbtxt')NUM_CLASSES = 90
The notebook code downloads and uses a pre-trained object detection model, ssd_mobilenet_v1_coco_2017_11_17 (built with the SSD method, which we talked briefly in the previous section, on top of the MobileNet CNN model, which we covered in the previous chapter). A ...
