Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jomit/sign-langauge-recognition
Sign Language Gesture Recognition on the Intelligent Edge using Azure Cognitive Services
https://github.com/jomit/sign-langauge-recognition
android-app azure cognitive-services
Last synced: about 2 months ago
JSON representation
Sign Language Gesture Recognition on the Intelligent Edge using Azure Cognitive Services
- Host: GitHub
- URL: https://github.com/jomit/sign-langauge-recognition
- Owner: jomit
- Created: 2018-06-19T01:47:09.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2023-03-24T23:14:46.000Z (almost 2 years ago)
- Last Synced: 2024-04-28T01:45:11.480Z (9 months ago)
- Topics: android-app, azure, cognitive-services
- Language: Java
- Size: 80.9 MB
- Stars: 12
- Watchers: 2
- Forks: 5
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Sign Language Gesture Recognition on the Intelligent Edge using Azure Cognitive Services
For this walkthrough we will use an Android Phone/Tablet as the Intelligent Edge device. The goal is to show how we can quickly create image recognition models using [Custom Vision Service](https://www.customvision.ai/) and export it to consume it offline at the Edge.
If you just want to test the app you can download the APK file from [here](https://github.com/jomit/sign-langauge-recognition/blob/master/sign-recognizer.apk?raw=true).
#### Setup
- [Install Android Studio](https://developer.android.com/studio/index.html)
- Download [dataset.zip](https://github.com/jomit/sign-langauge-recognition/blob/master/dataset.zip?raw=true) file and extract it
- Original dataset can be found [here](https://www.kaggle.com/datamunge/sign-language-mnist/version/1)
#### Create Sign Language Recognition ML Model
- Sigin to [Custom Vision Service](https://www.customvision.ai/) using your Azure Account
- Create New Project with Domains as **General (compact)**
- Upload all images from **dataset\A** folder with Tag **A**
- ***Repeat** above step for all alphabets in the dataset...*
- Click the **Train** button at top to start training the model
- Once the training is complete use the **Quick Test** button to upload a new image and test it.
#### Export the ML Model
- Under **Performance** tab click **Export**
- Select **Android (Tensorflow)** and download the zip file
- Extract the zip file and verify that it contains **model.pb** and **labels.txt** file
#### Create the Android App
- Clone [cognitive-services-android-customvision-sample](https://github.com/Azure-Samples/cognitive-services-android-customvision-sample) repo as a template or you can use the [android-mobileapp](https://github.com/jomit/sign-langauge-recognition/tree/master/android-mobileapp) code from this repo.
- Replace both **model.pb** and **labels.txt** files in `app\src\main\assets\`
- Open the project in Android Studio
- *Make any updates to UI / Labels as necessary*
#### Deploy the Android App on the device
- First enable Developer Mode + USB Debugging on the Android device
- See instructions for Samsung Galaxy S7 [here](https://www.androidcentral.com/how-enable-developer-mode-galaxy-s7)
- Connect your device to laptop via USB
- Click **Run** and select the **app**
- Select the **Connected Device**
- For first time you need to allow the camera and other permissions and run it again.
#### Testing
