Cloud vision api face detection no image present error - google-cloud-vision

Question

We are getting No image present. error while attempting face detection with the cloud vision api.
We are using code from the official documentation.
Please see the code below.
const request1={
"requests":[
{
"image":{
"content": imgdatauri //It contains image data uri
},
"features": [
{
"type":"FACE_DETECTION",
"maxResults":1
}
]
}
]
};
client
.annotateImage(request1)
.then(response => {
console.log(response);
response.send(response);
})
.catch(err => {
console.error(err);
response.send(err);
});
Here is the error message.
Error: No image present.
at _coerceRequest (/rbd/pnpm-volume/e40024d2-3d05-4f3d-a435-6d4e6ca96fb0/node_modules/.registry.npmjs.org/#google-cloud/vision/1.1.3/node_modules/#google-cloud/vision/src/helpers.js:69:21)
at ImageAnnotatorClient.<anonymous> (/rbd/pnpm-volume/e40024d2-3d05-4f3d-a435-6d4e6ca96fb0/node_modules/.registry.npmjs.org/#google-cloud/vision/1.1.3/node_modules/#google-cloud/vision/src/helpers.js:224:12)
at PromiseCtor (/rbd/pnpm-volume/e40024d2-3d05-4f3d-a435-6d4e6ca96fb0/node_modules/.registry.npmjs.org/#google-cloud/promisify/1.0.2/node_modules/#google-cloud/promisify/build/src/index.js:71:28)
at new Promise (<anonymous>)
at ImageAnnotatorClient.wrapper [as annotateImage] (/rbd/pnpm-volume/e40024d2-3d05-4f3d-a435-6d4e6ca96fb0/node_modules/.registry.npmjs.org/#google-cloud/promisify/1.0.2/node_modules/#google-cloud/promisify/build/src/index.js:56:16)
We would like to know what we need to do to resolve the issue.

Answer 1

Method 1:
In case of vision API, if image is stored locally you must convert that image to base64 string. Now this converted string is passed as value to content.
Make sure that you are converting image to base64 string and then passing to the content value.
There are some services available online to convert image to base64 string. You can also convert image to base64 by writing a piece of code. You can find services online and select anyone them. I am providing link to one service.
https://www.browserling.com/tools/image-to-base64
Method 2:
You can provide public url of image to vision API.
{
"requests":[
{
"image":{
"source":{
"imageUri": PUBLIC_URL
}
},
"features":[
{
"type":TYPE_OF_DETECTION,
"maxResults":MAX_NUMBER_OF_RESULTS
}
]
}
]
}
Method 3:
You can create bucket and put image there.
Now you can provide this image object URL or path.
I think this will help you.
Thank you.

Answer 2

I have created a Cloud Function with Cloud Storage Trigger, my function gets triggered ( with event ) when I upload an image file I can see there is event.mediaLink event.selfLink, I tried using both to load image, but it keeps complaining about No Image being present
here is the code
exports.analyzeImage = function(event) {
const vision = require('#google-cloud/vision');
const client = new vision.ImageAnnotatorClient();
console.log('Event', event.mediaLink)
let image = {
source: {imageUri: event.mediaLink}
};
return client.labelDetection(image).then(resp => {
let labels = resp[0].labelAnnotations.map( l => {
return {
description: l.description,
score: l.score
};
});
return labels;
// const dataset = bigquery.dataset('dataset')
// return dataset.table
}).catch(err => {
console.error(err)
})
}