iOS-Experiments Contents

Table of Contents of iOS-Experiments repository.

How to access to iOS-Experiments?

Here are some way to get access to the private repo:

• Become a sponsor (premium or higher): https://github.com/sponsors/shu223
• Let me know that. Here is my email address.

Table of Contents

Google Cloud Speech API sample

• A sample of Google Cloud Speech-to-Text API
• Modified ios-docs-samples/speech/Objective-C/Speech-gRPC-Nonstreaming in the official repository to recognize the audio of audio files.
  • 公式リポジトリのios-docs-samples/speech/Objective-C/Speech-gRPC-Nonstreamingを改変し、オーディオファイルの音声を認識するようにしたもの
• You should replace API_KEY when you use it.
  • 利用にあたってはAPI_KEYを差し替えること

vottcsv2createml

• Tool to convert annotation files as VoTT CSV format to Create ML JSON format.
  • VoTT CSVフォーマットのアノテーションファイルをCreate MLのObject Detection用のJSONフォーマットに変換するツール
• I made this for myself, so it's not very kind and incomplete.
  • 自分用につくったのでいろいろ不親切＆不完全

yolo2createml

• Tool to convert annotation files as YOLO format (.txt) in YOLO format to Create ML JSON format.
  • YOLOフォーマットのアノテーションファイル（.txt）をCreate MLのObject Detection用のJSONフォーマットに変換するツール
• I made this for myself, so it's not very kind and incomplete (for example, labels are hard coded)
  • 自分用につくったのでいろいろ不親切＆不完全（ラベルがハードコーディングされてたり）

ClockGestureClassification

Activity Classification demo on watchOS. The watchOS app can classify clockwise gesture and counter-clockwise gesture.

Here is the demo video: youtu.be/I09hUyz1oG8?t=1939

• Source code: iOS-Experiments/ClockGestureClassification
  • 学習データ収集用に作成したwatchOSアプリ、学習データ、Create MLプロジェクトは含んでいません

watchOS, Activity Classification, Machine Learning, Deep Learning, Core ML, Create ML

https://github.com/shu223/iOS-Experiments/tree/master/

Metal on watchOS

Metal shader examples on watchOS. Though Metal framework is NOT supported on watchOS, SceneKit framework and SCNShadable protocol is supported and it actually works! It contains toon shading sample and grayscale sample written in MSL (Metal Shading Language).

• Source code: iOS-Experiments/watchOS-Metal

watchOS, Metal, Metal Shading Language, shader modifier, SceneKit, MSL, GPU

AnimeSpeakerClassifier

An experimental app to classify speakers in an anime movie using SoundAnalysis framework. It worked well. Used Create ML to train the model.

SoundAnalysisフレームワークを用いてアニメ動画の話者を識別できるか試した。かなりうまくいった。学習はCreate MLを使用。

• Source code: iOS-Experiments/AnimeSpeakerClassifier
  • The Core ML model and video data have been removed from the project in consideration of copyright, so you can NOT build & run this project.
  • 著作権に配慮してCore MLモデルと動画データをプロジェクトから除去しているのでビルド不可
  • The Create ML project file to train the model and the training data are NOT contained in the repo.
  • Create MLのプロジェクトと学習用の音声データも同梱していません

SoundAnalysis, Speaker Classification, Video Processing, iOS 13

PodcastSpeakerClassifier

An experimental app to classify speakers in a podcast audio using SoundAnalysis framework. It did NOT well. Used Create ML to train the model.

SoundAnalysisフレームワークを用いてPodcastの話者を識別できるか試した。あまりうまくいってない。学習用の音声データが少なかったかもしれない。学習はCreate MLを使用。

• Source code: iOS-Experiments/PodcastSpeakerClassifier
  • The Create ML project file to train the model and the training data are NOT contained in the repo.
  • Create MLのプロジェクトと学習用の音声データは同梱していない。どこかにあるはずなので見つけたらアップするかも。

SoundAnalysis, Sound Analysis, Speaker Classification, Audio Processing, iOS 13

Scene Reconstruction with SceneKit + ARKit

Visualizing ARMeshGeometry from LiDAR without RealityKit (with SceneKit)

• Source code: iOS-Experiments/ARKit_SceneReconstruction

ARKit 3.5, LiDAR, Depth, iOS 13.4

Core Haptics samples

HapticsMaker

A sample to generate custom haptic patterns using Core Haptics framework which is added in iOS 13.

• Source code: iOS-Experiments/CoreHaptics/HapticsMaker

AHAPPlayer

A very simple example to play an AHAP file.

• Source code: iOS-Experiments/CoreHaptics/AHAPPlayer

CoreHaptics, Core Haptics, iOS 13

MSL from GLSL

MSL (Metal Shader Language) samples which are ported from GLSL.

• Source code: iOS-Experiments/MSLfromGLSL/MSLfromGLSL1
• Original: http://glslsandbox.com/e#36694.0

• Source code: iOS-Experiments/MSLfromGLSL/MSLfromGLSL2
• Original: http://glslsandbox.com/e#36614.0

• Source code: iOS-Experiments/MSLfromGLSL/MSLfromGLSL3
• Original: http://glslsandbox.com/e#36538.3

Metal, Metal Shading Language, MSL, GPU

Rectangle Detection & Correction

A real-time camera sample to detect a rectangle using Vision (VNDetectRectanglesRequest, VNRectangleObservation) and correct it using Core Image (CIPerspectiveCorrection).

• Source code: iOS-Experiments/RectangleDetection

Vision, Core Image, Image Processing

PhysicalFaceTracker

An one-day project to make a face tracking camera using iPhone and Raspberry Pi (bleno).

• Source code: iOS-Experiments/PhysicalFaceTracker
  • iOS code only. The code on Raspberry Pi and the 3D model data of the hardware are NOT included.

Vision, CoreBluetooth, Image Processing, IoT, Hardware

Processing with Swift

Solar System 2D

Original(Processing): https://youtu.be/l8SiJ-RmeHU

• Source code: iOS-Experiments/Processing/SolarSystem2D

SpriteKit, Animation

Solar System 3D

Original(Processing): https://github.com/CodingTrain/website/tree/master/CodingChallenges/CC_008_SolarSystemGenerator3D

• Source code: iOS-Experiments/Processing/SolarSystem3D

SceneKit, Animation, 3D

ZundokoBLE

• https://twitter.com/shu223/status/715946579585314820
• https://qiita.com/shu223/items/2c7b2e1ed058a1e3b33f

• Source code: iOS-Experiments/ZundokoBLE
  • It can NOT be built with the latest Xcode because written in Swift 3.x.

CoreBluetooth, Core Bluetooth, BLE