Skip to content

Latest commit

 

History

History
34 lines (23 loc) · 1.25 KB

README.md

File metadata and controls

34 lines (23 loc) · 1.25 KB

CornConeDetector

This application can detect corn (sweet corn) or cone (road cones, traffic cones or pylons). The character on the left of the app icon is すみっコぐらし (Sumikko Gurashi)'s こーん (corn).

Corn Cone こーん
Corn Cone こーん

It is base on the live capture sample code provided by Apple: Recognizing Objects in Live Capture.

Requirements

  • iOS 14.0+
  • Xcode 12.0+

Model training

The object detection model used in this app (CornConeDetector.mlmodel) was trained by Create ML's Object Detection templete. Thanks to transfer learning, the model size is small (6.8 MB). The model uses Vision Feature Print, which is a feature extractor optimized for object detection.

Parameters

  • Algorithm: Transfer Learning
  • Iterations: 300
  • Batch Size: Auto

Learning curve