ColorPredictor

Takes a full path to an image and returns its predicted color. Will be integrated to work with live cam in order to detect colors of objects in real time. This project uses Node.js with integrated ECMAScript. So it's essential to run your code with node -r esm your-file-name.js Or you can edit this project in a way there wouldn't be a need for esm. Your choice.

How To Use

As npm package

Download with npm install color-predictor or yarn add color-predictor Request or import package and call predictColor('path/to/img'). Sample code:

var predictColor = require("color-predictor")
var path = "../../purple.jpg"
predictColor(path)

Run your code with node -r esm your-file-name.js

As a seperate project

IMPORTANT

As I've adjusted the code for npm package, the path in the project where it calls for the model.json is different than if you had done npm install. Right now it is :

in predict.js: var path = "./node_modules/color-predictor/my-model/model.json"

But if you are using this option, you should change this to :

in predict.js: var path = "./my-model/model.json

or wherever your model.json is located.

Downloading to same directory of your project

Download this repository to the same directory of your project either manually or with Git. Run npm install Example usage:

var predictColor = require("./color-finder")

var path = "path/to/image.jpg"
predictColor(path)

and run node -r esm your-file-name.js

Since this project uses esm with node, you can either use require or import. Both works.

Downloading to a seperate directory

Download this repository either manually or with Git. Copy and paste package.jsonto your working directory, or copy the dependencies inside package.json and add them to your own project's package.json. Run npm install. Example usage:

var predictColor = require("path/to/color-classification/color-finder")

var path = "path/to/image.jpg"
predictColor(path)

and run node -r esm {fileNameToBeRun.js}

Since this project uses esm with node, you can either use require or import. Both works.

Note

The path to the trained model is given in predict.js. If you would like to use another model, just require it.