Amazon Rekognition announces updates to its face detection, analysis, and recognition capabilities
Today we are announcing updates to our face detection, analysis, and recognition features, providing customers with improvements in the ability to detect more faces from images, perform higher accuracy face matches, and obtain improved age, gender, and emotion attributes for faces in images. Amazon Rekognition customers can use each of these enhancements starting today, at no additional cost. No Machine Learning experience is required.
Today we are announcing updates to our face detection, analysis, and recognition features, providing customers with improvements in the ability to detect more faces from images, perform higher accuracy face matches, and obtain improved age, gender, and emotion attributes for faces in images. Amazon Rekognition customers can use each of these enhancements starting today, at no additional cost. No Machine Learning experience is required.
“Face detection” tries to answer the question: is there a face in this picture? In real world images, various aspects can impact a system’s ability to detect faces with high accuracy: pose variations caused by head movement and/or camera movements, occlusion due to foreground or background objects (faces covered by hats, hair, hands, or another person), illumination variations such as low contrast and shadows, or bright lighting leading to washed out faces, low quality and resolution leading to noisy and blurry faces, and distortion from cameras and lenses themselves. These issues manifest as both missed detections (a face not detected) or false detections (an image region detected as face even when there is no face). For example, in social media different poses, camera filters, lighting and occlusions (photo-bomb) are common. For financial services customers, verification of customer identity as a part of multi-factor authentication and fraud prevention workflows involves matching a high resolution selfie (a face image) with lower resolution, small and often blurry image of face on a photo identity document (such as a passport or driving license). Also, many customers have to detect and recognize faces of low contrast from images where the camera is pointing at a bright light.
With the latest updates, Amazon Rekognition can now detect 40% more faces - that were previously missed - in images that have some of the most challenging conditions stated above, while reducing the rate of false detections by 50%. This means that customers such as social media apps can get consistent and reliable detections (fewer misses, fewer false detections) at higher confidence, allowing them to deliver better customer experiences in use cases like automated profile photo review. In addition, face recognition now returns 30% more correct 'best' matches (the most similar face) compared to our previous model when searching against a large collection of faces, enabling customers to get better search results in applications like fraud prevention. Face matches now also have more consistent similarity scores across varying lighting, pose and appearance, allowing customers to use higher confidence thresholds, avoid false matches, and reduce human review in applications such as identity verification. As always, for use cases involving civil liberties or customer sentiments, where the veracity of the match is critical, we recommend customers use best practices, higher confidence level (at least 99%), and always include human review.
Amazon Rekognition face detection and face recognition enhancements are available today in all regions where Amazon Rekognition Image and Video are offered. You can get started today via the Amazon Rekognition Console or by downloading the latest AWS SDK. For more information please refer to our updated documentation.