tezvyn:

Recognize objects in ARKit with a Core ML model

Source: interviewadvanced

WHAT IT TESTS: Combining ARKit, Vision, and Core ML plus 2D-to-3D mapping. OUTLINE: Pull the frame's capturedImage, run Vision/Core ML off the main thread, raycast the 2D detection point into the scene, add an ARAnchor.

WHAT IT TESTS: Whether you can architect a real-time ARKit plus ML pipeline and handle the 2D-to-3D leap. ANSWER OUTLINE: In the session delegate read ARFrame.capturedImage (a CVPixelBuffer), run a VNCoreMLRequest on a background queue so the 60fps render loop is not blocked, and apply the correct orientation. Map the 2D bounding-box center to a world position with raycasting (raycastQuery) against detected planes, then add an ARAnchor.

Read the original → interview

Get five bites like this every day.

Tezvyn delivers a daily feed of 60-second tech bites with quizzes to lock in what you learn.

Recognize objects in ARKit with a Core ML model · Tezvyn