如何在 javascript 中为 tensorflow 加载本地图像

Question

I want to load a data.jpg file to be used by tesnorflow mobilenet but if i require the the file normally i get So my question is how would i load an image in my javascript code to be used i have posted all the methods i tried Requiring normally

node .
/home/arnav7633/jarvis/data.jpg:1
����


SyntaxError: Invalid or unexpected token
   at Module._compile (internal/modules/cjs/loader.js:703:23)
   at Object.Module._extensions..js (internal/modules/cjs/loader.js:770:10)
   at Module.load (internal/modules/cjs/loader.js:628:32)
   at Function.Module._load (internal/modules/cjs/loader.js:555:12)
   at Module.require (internal/modules/cjs/loader.js:666:19)
   at require (internal/modules/cjs/helpers.js:16:16)
   at Object.<anonymous> (/home/arnav7633/jarvis/index.js:2:13)
   at Module._compile (internal/modules/cjs/loader.js:759:30)
   at Object.Module._extensions..js (internal/modules/cjs/loader.js:770:10)
   at Module.load (internal/modules/cjs/loader.js:628:32)
   ```
   If i use fs for readstream i get 
   ```
(node:20025) UnhandledPromiseRejectionWarning: Error: pixels passed to tf.browser.fromPixels() must be either an HTMLVideoElement, HTMLImageElement, HTMLCanvasElement, ImageData in browser, or OffscreenCanvas, ImageData in webworker or {data: Uint32Array, width: number, height: number}, but was ReadStream

If i use an url

    (node:20149) UnhandledPromiseRejectionWarning: Error: pixels passed to tf.browser.fromPixels() must be either an HTMLVideoElement, HTMLImageElement, HTMLCanvasElement, ImageData in browser, or OffscreenCanvas, ImageData in webworker or {data: Uint32Array, width: number, height: number}, but was String

Here is my code 

    const mobilenet = require('@tensorflow-models/mobilenet');
const img = 'https://imgur.com/a/IPaQnZ5'
require('@tensorflow/tfjs-node')
const func = async(img) => {
console.log('called')

// Load the model.
const model = await mobilenet.load();
console.log('loaded')
// Classify the image.
const predictions = await model.classify(img);

console.log('Predictions: ');
console.log(predictions);
}
(async() => func(img))()

Here is the code in the docs

const mobilenet = require('@tensorflow-models/mobilenet');

const img = document.getElementById('img');

// Load the model.
const model = await mobilenet.load();

// Classify the image.
const predictions = await model.classify(img);

console.log('Predictions: ');
console.log(predictions);

Link to therepo

Answer 1

The code of the docs added to the question is for a prediction using the browser.

Here is how to load an image file and use it for prediction with nodeJs

const tfnode = require('@tensorflow/tfjs-node')

const imageBuffer = await fs.readFile(file_path);
const tensor = tfnode.node.decodeImage(imageBuffer)
...
const predictions = await model.classify(tensor);