1

I'm using the gcloud Node SDK to detect text in small images. It works ok for most of the samples but then, there are a couple of rather obvious image samples that somehow GCloud fails to read.

However when i try the exact image on Vision landing page GCloud Vision it works just perfectly.

    const vision = require('@google-cloud/vision')
    vision.detectText('my-image.jpg').then((result) => {
       // result is empty
    });

Sample Image:

enter image description here

Gcloud Node SDK Result (EMPTY):

[[],{"responses":[{"faceAnnotations":[],"landmarkAnnotations":[],"logoAnnotations":[],"labelAnnotations":[],"textAnnotations":[],"fullTextAnnotation":null,"safeSearchAnnotation":null,"imagePropertiesAnnotation":null,"cropHintsAnnotation":null,"webDetection":null,"error":null}]}]

GCloud Vision Landing page demo result (CORRECT) :

enter image description here

maephisto
  • 4,952
  • 11
  • 53
  • 73
  • I'm using the png file that you shared and I don't get any `Text` results on the landing page. Did you use a different file format when you first tried it? – user835611 Mar 02 '19 at 11:51

0 Answers0