/InstanceSegmentation-CoreML

Primary LanguageSwiftApache License 2.0Apache-2.0

InstanceSegmentation-CoreML

What is instance segmentation?

image source: https://images.app.goo.gl/tqCZU7uG7WfWwAt19

Features

  • Inference the model and perform post-process yolact model
  • Import a photo from device's photo library, inference, and draw the result on the screen
  • Capture every frame from device's camera, inference, and draw the result on the screen
  • Make it real-time when capture from camera

Requirements

Getting Started

Models

DEMO reference: https://github.com/Ma-Dan/Yolact-CoreML

Model size, minimum iOS version, input/output shape

Model Size (MB) Minimum iOS Version Input Shape Output Shape Download Link Source Link
yolact.mlmodel 146.9 iOS11.2 [1, 550, 550, 3] 4 MLMultiArrays Link link

Inference time (ms)

Total time (ms)

FPS

How it works

Project Structure

InstanceSegmentation-CoreML
├── InstanceSegmentation-CoreML
|   ├── AppDelegate.swift
|   ├── Assets.xcassets
|   ├── Base.lproj
|   ├── Info.plist
|   ├── mlmodels
|   |   └── yalact.mlmodel				# you need to download from release of this repository
|   └── ViewController.swift
├── InstanceSegmentation-CoreML.xcodeproj
├── README.md
└── LICENSE

Models & Papers

References