1.6 Image Classification Algorithm Example

The mobilenetv2.py program is installed on the development board for testing the image classification algorithm functionality of MobileNet V2. This program reads the static image zebra_cls.jpg as the model input and outputs the classification results on the command-line terminal.

Execute the mobilenetv2.py program:

sunrise@ubuntu:~$ cd /app/pydev_demo/01_classification_sample/02_mobilenetv2/
sunrise@ubuntu:/app/pydev_demo/01_classification_sample/02_mobilenetv2$ sudo python3 ./mobilenetv2.py

Expected Output

The prediction results of the image classification algorithm are output. zebra_cls.jpg is an image of a zebra. According to the ImageNet dataset classification, the returned result shows that the confidence for "zebra" is 0.9936.

=== Scheduling Parameters ===
mobilenetv2_224x224_nv12:
  priority    : 0
  customId    : 0
  bpu_cores   : [0]
  deviceId    : 0
Top-5 Predictions:
zebra: 0.9936
tiger, Panthera tigris: 0.0039
hartebeest: 0.0007
tiger cat: 0.0007
impala, Aepyceros melampus: 0.0003