PinHead is an Angular2 app based on mgechev's seed project that repins based on your face's emotion captured by webcam and analyzed by Google's Vision API. This project is my active exploration of AngularJS and RxJS.
A little while ago I set out on a journey learning about this extremely sweet and clever kitten that is Angular 2. Not too long ago Google released its Vision API sentiment analysis. I wanted to focus on Angular 2 with RxJS, so a server-less backend was a convenient way to avoid scope creep. I took a brief detour to set up a rudimentary server-less API endpoint that analyzes an image for facial sentiment. For more details on this, check out the gist on Face Sentiments with Google Vision API via AWS API Gateway and Lambda.
You can check out the latest version by visiting pin-head.cyk.im — it requires camera access and a Pinterest account with boards (owned and followed).
After choosing a board to pin from and a board to pin to, a pin will be presented every few seconds. Then, behind the scenes...
- A snapshot from your browser camera is taken
- The snapshot is encoded and sent off to a sentiment endpoint
- The Google Vision API then analyzes the image for faces and their sentiments
- If the likeliest sentiment is JOY then the related pin in pinned to you board.
Obviously, it works best with kittens and puppies. 😉
You'll need a Pinterest App and the Google Vision API via AWS API Gateway and Lambda. Once you have a Pinterest app id and a sentiments endpoint, update the project.config.ts file.
$ npm install
$ npm start
Please refer to the mgechev/angular2-seed documentation for additional setup and usage instructions.
Frameworks, libraries, and web APIs used:
- Angular 2 (mgechev/angular2-seed)
- Angular 2 Material
- RxJS 5
- Pinterest API
- Google Cloud Vision API (with a bit of AWS API Gateway and Lambda)
Pinhead icon by Icons8
MIT