EITC/AI/GVAPI Google Vision API

Google Cloud offers two computer vision services (jointly referred to as Vision AI) that use machine learning to understand images and videos with high prediction accuracy, i.e. AutoML Vision and the Vision API. The AutoML Vision automates training of the custom machine learning models. It enables upload of images and training custom image models with easy-to-use graphical interface; optimizing the models for accuracy, latency, and size and exporting them to any application in the cloud, or to an array of devices at the edge. On the other hand Google Cloud’s Vision API offers powerful pre-trained machine learning models through REST (Representational State Transfer) and RPC (Remote Procedure Call) APIs, assigning labels to images and quickly classifying them into millions of predefined categories. Detecting objects and faces, reading printed and handwritten text, and building valuable metadata into image catalogs. You can thus use AutoML Vision to derive insights from images in the cloud or at the edge or use the pre-trained Vision API models to detect emotion, understand text from visual data, and more.

With Google Cloud’s Vision API it is possible to:

• Detect objects: Detect objects, where they are, and how many.
• Enable vision product search: Compare photos to images in your product catalog, and return a ranked list of similar items.
• Detect printed and handwritten text: Use OCR and automatically identify language.
• Detect faces: Detect faces and facial attributes. (Face recognition not supported.)
• Identify popular places and product logos: Automatically identify well-known landmarks and product logos.
• Assign general image attributes: Detect general attributes and appropriate crop hints.
• Detect web entities and pages: Find news events, logos, and similar images on the web.
• Moderate content: Detect explicit content (adult, violent, etc.) within images.
• Celebrity recognition: Identify celebrity faces in images (limited access, see documentation.)
• Classify images using predefined labels: Pre-trained models leverage vast libraries of predefined labels.
• Use Google’s data labeling service: Google can help annotating images, videos, and text.
• Use APIs: Use REST and RPC APIs.

The possible use cases for Vision API are countless.

For example using Vision API you can implement vision product search enabling your customers to find products of interest within images and visually search product catalogs (image search, automatically designated similar products, etc.).

The video above explains how Google’s Cloud AutoML Vision uses AI to analyze images.

A twin AI system, closely related to the pre-trained and constantly upgraded Google Vision API is Google AutoML Vision enabling enterprises to use their own machine learning models and custom training for the artificial intelligence assistance in vision analysis and understanding. Part of Google Cloud’s machine learning suite of products, it’s designed to help developers with limited machine learning expertise train custom vision models for their specific use cases. In need of on-demand access to the general pre-trained model, AI developers hould use the Google Vision API.