AlchemyVision Face Detection/Recognition API Identifies Celebrities Using Data From Knowledge Graph

AlchemyAPI, a leading advanced natural language processing and computer vision technology company, has announced the public release of the AlchemyVision Face Detection and Recognition API, which is capable of detecting faces in photos and returning information such as position, age and gender. In addition, the API is able to identify the faces of many celebrities and return name, gender, type hierarchy information, website and other metadata pulled from AlchemyAPI's custom-built knowledge graph.


The AlchemyVision Face Detection and Recognition API demo was able to recognize and identify actors from the Vampire Diaries television series. View full-size image. Image credit: Google Images

The AlchemyVision Face Detection and Recognition API can provide the approximate age for faces identified in photos as well as their gender. The API will automatically attempt to recognize and identify faces in photos, matching them to the thousands of celebrity faces stored in the proprietary knowledge graph developed by AlchemyAPI. The knowledge graph was released with more than 60,000 celebrity faces, and thousands more have been added since the release date. The API can return celebrity metadata, which includes knowledge graph hierarchy information (people, celebrities, etc.), subtypes (Person, Actor, FilmActor, TVActor, etc.) and linked data (dbpedia, freebase, yago, etc.).

ProgrammableWeb reached out to AlchemyAPI CMO Richard Leavitt to find out more about the new API and the company's knowledge graph. Leavitt explained that the knowledge graph containing the celebrity metadata returned by the API is custom coded and was built from the ground up. He also said that this is the first time pathways of the knowledge graph have been allowed to be exposed via API.


The AlchemyVision Face Detection and Recognition API demo did not recognize two of the three actors from The Originals television series. However, the Platform/API can be taught to recognize and identify these faces. View full-size image. Image credit: Google Images

At the time of this writing, the AlchemyVision API featured image tagging, image search and facial-recognition capabilities. Leavitt told ProgrammableWeb that the company is working on two more AlchemyVision APIs that will be publicly released in the near future. One of the APIs will be capable of recognizing and extracting text from within images. For example, the API will be able to extract text from street signs shown in an image. The other API in development will be able to recognize and identify company and brand logos (logos that do not contain words). Leavitt also said that the company is developing an API that can recognize scenes in images such as mountains and beaches.

AlchemyAPI is working on a pretty ambitious road map that not only includes two new upcoming AlchemyVision APIs, but also a semantic data store service called AlchemyData and an answers API that will work somewhat like IBM Watson.

For more information about the new AlchemyVision Face Detection and Recognition API, visit the AlchemyAPI website.

See other ProgrammableWeb articles featuring AlchemyAPI, including an article about the launch of AlchemyVision, an article about how five natural language processing APIs Stack up and an article about the rapid rise of computer vision technology.

Be sure to read the next Machine Learning article: API Puts Machine Learning Functionality at Developers' Fingertips