COMP7506 2A Group Project

Group36, Members: CHEN Lu, WANG Junheng, WANG Yifei, LIU Wanting

If you have any question, please reach out to link@connect.hku.hk.

Build

This application is build by Xcode version 14.3 (14E222b) on macOS 13.3.1 (22E261).

Before building this application, make sure you download YOLOv3Tiny, PoseNetMobileNet100S16FP16, and MobileNetV2. Clicking links will directly download.

Drag each download to each model directory:

1. YOLOv3Tiny: ./MobileVision/YOLO

2. PoseNetMobileNet100S16FP16: ./MobileVision/PoseNet/Model

3. MobileNetV2: ./MobileVision/MobileNet/Image Predictor

If you wish to use different models, you can download from here. Do remember to change the model name corresponding to your downloads:

1. YOLO: Line24 in ./MobileVision/YOLO/VisionObjectRecognitionViewController.swift

2. PoseNet: Line44 in ./MobileVision/PoseNet/Model/PoseNet.swift

3. MobileNet: Line27 in ./MobileVision/MobileNet/Image%20Predictor/ImagePredictor.swift

After importing the models, build this project in Xcode on a physical iOS device because it has to use the camera.

Acknowledgements

1. Object Detection example is modified from Apple 'Recognizing Objects in Live Capture'.

2. Pose Estimation example is modified from Apple 'Detecting Human Body Poses in an Image'.

3. Image Classification example is modified from Apple 'Classifying Images with Vision and Core ML'.