Amazon Rekognition Now Offers Face Recognition And Text In Image Recognition

Amazon Rekognition, a part of Amazon Web Services, has added three new features to its AWS Service offerings. Now, users can recognize and extract the text in images, real-time face recognition, and more precise face detection in challenging crowded photos.

Via a detailed blog post, the company has informed that the customers will now experience up to 10 percent more accuracy in face verification and identification. Well, that’s quite an improvement.

The Text in Image recognition will be quite a useful for those who want to extract the text embedded in images, like license plates, street signs, news captions etc. The official announcement describes,

“You can use Rekognition Text in Image to recognize and extract textual content from images. Text in Image is specifically built to work with real-world images rather than document images. It supports text in most Latin scripts and numbers embedded in a large variety of layouts, fonts, and styles, and overlaid on background objects at various orientation as banners and posters.”

With the help of this amazing AI, the customers will be able to run visual search on social media applications; categorize videos based on relevant text on screen, such as - ads, news, sport scores, and captions; and in the images captured by street cameras, they can identify vehicles based on license plate numbers.

A real-time face recognition from a collection of tens of millions of faces is another great offering from Amazon Rekognition. For this, the company tells,

"This represents a 5-10X reduction in search latency, while simultaneously allowing for collections that can store 10-20X more faces than before. In security and safety applications, you can now identify people of interest against a collection of millions of faces in near real-time, enabling use cases such as timely and accurate crime prevention."

The third feature is not actually a new feature but an expansion of its old service of face identification. Now, users can also detect, analyze, and index up to 100 faces (which was previously 15) in a single image. The company praises this feature stating,

"With this improvement, you can accurately capture demographics and analyze the sentiments for all faces in group photos, crowded events and public locations such as airports and department stores." 

If you are interested to try out these features, you need to download the latest SDK or simply log in to the Amazon Rekognition Console. You can use sample images or upload your own for getting started. More information can be found at Amazon Rekognition documentation.