rocketkind
AI

Google Cloud Vision

Google Cloud Vision is Google Cloud's pre-trained image-analysis API: labels, text (OCR), faces, landmarks, logos, objects, explicit-content screening and reverse image search over a single image, plus OCR over PDF, TIFF and GIF files.

Authentication

  • Service account (JWT bearer)

Actions

Generic API request. Call any Google Cloud Vision API endpoint on this connection, for the case where no shipped action covers it yet — Product Search (product sets, products, reference images) is reached this way. The connection supplies the base URL, the authentication and the egress route, so only the path, query and body are yours…

  • Annotate file (PDF/TIFF, up to 5 pages)Run Vision detections — almost always DOCUMENTTEXTDETECTION — over up to five pages of a PDF, TIFF or GIF and get the results back in this call.
  • Annotate imageRun several Vision detections on one image in a single call — for example labels, text and safe search together — choosing each feature and its result cap.
  • Detect document text (dense OCR)Read a dense document image — a scanned page, receipt, form or handwriting — with the OCR model tuned for documents, returning the text with its page, block, paragraph and word structure plus confidence scores.
  • Detect explicit content (safe search)Screen an image for adult, violent, racy, medical or spoofed content before you publish or store it.
  • Detect facesFind the faces in an image with their position, head angles, facial landmarks and emotion likelihoods (joy, sorrow, anger, surprise).
  • Detect labelsLabel an image with what it shows — objects, scenes, activities, materials — each with a confidence score.
  • Detect landmarksRecognise well-known natural and man-made landmarks in an image and where they are on the map.
  • Detect logosRecognise popular product and brand logos in an image, for brand monitoring or catalog tagging.
  • Detect text (OCR)Read the text in a photo — signs, labels, screenshots, short snippets in a scene — with optical character recognition.
  • Detect web entities and pagesSearch the web for an image: the entities it is associated with, exact and partial copies of it, visually similar images, the pages it appears on and Google's best-guess label.
  • Get operationCheck on a long-running Vision job started by googlecloudvision:asyncannotatefile: whether it is still running, and once done, where its results were written or why it failed.
  • Localize objectsFind each distinct object in an image and where it is — a box per object, unlike labels, which describe the image as a whole.
  • Start async file annotation (PDF/TIFF to Cloud Storage)Start OCR (or any other Vision detection) over a whole PDF, TIFF or GIF in Cloud Storage, with no page limit; Vision writes the results as JSON files under a Cloud Storage prefix you choose.

Product names, logos and brands are the property of their respective owners and are used here for identification only. Their use does not imply endorsement by, or affiliation with, those owners.