BatchAnnotateImagesResponse images info (context support) for web based images

[dragan1982]

Hi all. How could one know in what exact web based image the text was detected when multiple web based images are sent to the cloud vision api in a single request using BatchAnnotateImagesRequest? BatchAnnotateImagesResponse doesn't return that information which is kinda odd... It has ImageAnnotationContext, which holds image details, but it's reserved only for files and not web based images.

Is there some way to do this? Maybe like preserving order of images in request / response or something down that line.

[abdelilahf]

Hi,

looking to our documentation [1] the batch annotation images response return an Individual responses to image annotation requests within the batch as AnnotateImageResponse [2].

ImageAnnotationContext, if present, contextual information is needed to understand where this image comes from.[1]https://cloud.google.com/vision/docs/reference/rpc/google.cloud.vision.v1#batchannotateimagesrespons...
[2]https://cloud.google.com/vision/docs/reference/rpc/google.cloud.vision.v1#annotateimageresponse

[dragan1982]

Hi, abdelilahf

Sorry to see response a bit late. I was on vacation.
As I wrote, last time we checked, contextual information was just implemented to return image information about file based images, not web based images, and we only can check batches of images fetched from image service via api "on the fly".

Do you maybe know if this will be implemented any time soon for web images? Is there some way to request this feature from Google? Also, is there some workaround in the meantime to cover our use case?

[dragan1982]

Solved with this answer. Thanks!