Model Inference - Inference on Cropped Regions

This tutorial shows how to inference model with Generic Inference with cropped regions. The crop configuration will do cropping image on device, execute inference only on the cropped areas of image, get the raw output from device, and does post-processing in the software.

The flow in concept:

  1. Setting crop information in kp.GenericRawImageHeader
  2. Send an image to inference
  3. Receive result N times (N specify for number of crop bounding boxes)

Note: Please upload NEF model on Kneron device before the following tutorial. See the Load NEF Model for details.

Reference Examples:


Inference on Cropped Regions

Recommend use cv2 to read the image or capture camera frame.