One week prior to the annual AWS re:Invent event at Nevada, the cloud player has packed a bunch of exciting additions to its face recognition product (Amazon Rekognition). The updates aim to provide machine learning-based computer vision capabilities to Amazon Web Services customers.
New feature updates to Amazon Rekognition
Detection and Recognition of Text in Images: From now, users can use Rekognition Text in Image to recognize and extract textual content from images. The new text in image feature is specifically designed to work with real-world images than document images. Extensive text support is available for numbers and Latin scripts that are embedded in a vast variety of layouts, fonts, and styles coming in various orientations such banners, posters and so on. The overwhelming demand from the AWS users on a provision to recognize text embedded in images, such as street signs and license plates captured by traffic cameras, news, and captions on TV screens, or stylized quotes overlaid on phone-captured images etc. has led the Amazon team to work on this.
Real-time face recognition across millions of faces: The next addition allows Amazon Rekognition users to perform real-time face searches against vast collections with tens of millions of faces. That means, the collections can now store 10-20x more faces than before. There is a commendable 5-10x reduction in the search latency. This feature will find its wider application in security and safety, crime suspects detection, and law enforcement cases. The Washington County Sheriff’s Office reports that Amazon Rekognition has helped them in reducing the suspect identification time from 3 days to a few minutes.
Detect up to 100 faces in challenging crowded photos: AWS Rekognition can now detect up to 100 different faces in a single photo, compared to its previous limit of 15 faces. With this feature, you can precisely capture demographics of all the faces detected, and analyze the emotions and sentiments for each face in the group picture. This will boost the security of crowded events and public locations such as airports and department stores.
The changes prove that Amazon Rekognition is keenly vying with Microsoft Azure Face API and other competitors such as IBM, Salesforce, and Google to secure its position as the cloud leader and also a key AI (Artificial Intelligence) player. With each one focusing to make machine learning revolution more accessible to customers, the world is expecting to receive better tools and technologies that promise intelligent results.