Alchemy API, known for its natural language processing technologies for text mining and semantic analysis, has launched a computer vision API for analyzing and tagging images.
AlchemyVision is a cloud-based service that uses a unsupervised learning to classify images and assign probabilities to multiple tags. The new platform combines AlchemyAPI's core NLP technology with 3D image analysis and, unlike other similar systems that are under development (by Google, for example) is available for general use.
You can try the AlchemyVision API using its demo, where you can use one of your own images or one in its gallery.
So far two APIs are available:
Image Tagging API: When provided an image file or URL, AlchemyVision returns between 0 and 20 keywords, summarizing scenes, objects, and stylistic features. The API is capable of identifying 3D objects such as chairs, dogs, buildings, and recognizing scenes such as streets, stores, beaches, landscapes, mountains, and detecting people, faces, etc.
Image Extraction API: Given any URL, the system performs a scan of the page, finds the most prominent image and retrieves the image file directly.
These are to be followed by APIs for image search, text extraction, and recognizing faces and logos.
There is a free tier that lets you analyse up to 1000 images per day after that you have to pay.