User:Hwd3400/Amazon Rekognition

Amazon Rekognition is a cloud-based Software as a service (SaaS) computer vision platform that was launched in 2016.

Capabilities
Rekognition provides a number of computer vision capabilities, which can be divided into two categories: Algorithms that are pre-trained on data collected by Amazon or its partners, and algorithms that a user can train on their own data.

As of July 2019, Rekognition provides the following computer vision capabilities.

Pre-trained algorithms

 * Celebrity recognition in images
 * Facial attribute detection in images, including gender, age range, emotions (e.g. happy, calm, disgusted), whether the face has a beard or mustache, whether the face has eyeglasses or sunglasses, whether the eyes are open, whether the mouth is open, whether the person is smiling, and the location of several markers such as the pupils and jaw line.
 * People Pathing enables tracking of people through a video. An advertised use-case of this capability is to track sports players for post-game analysis.
 * Text detection and classification in images
 * Unsafe visual content detection

Algorithms that a user can train on their own data

 * SearchFaces enables users to import a database of images with pre-labeled faces, to train a machine learning model on this database, and to expose the model as a cloud service with an API. Then, the user can post new images to the API and receive information about the faces in the image. The API can be used to expose a number of capabilities, including identifying faces of known people, comparing faces, and finding similar faces in a database.
 * Face-based user verification

Uses

 * In late 2017, the Washington County, Oregon Sheriff's Office began using Rekognition to identify suspects' faces. Rekognition was marketed as a general-purpose computer vision tool, and an engineer working for Washington County decided to use the tool for facial analysis of suspects.


 * In April 2018, it was reported that FamilySearch is using Rekognition to enable their users to "see which of their ancestors they most resemble based on family photographs."
 * In May 2018, it was reported that Orlando, Florida was running a pilot using Rekognition for facial analysis in law enforcement. That pilot ended in July 2019.
 * In September 2018, it was reported that Mapillary is using Rekognition to read the text on parking signs (e.g. no stopping, no parking, or specific parking hours) in cities.

Racial and gender bias
In 2018, researchers Joy Buolamwini and Timnit Gebru published a study called Gender Shades. In this study, a set of images was collected, and faces in the images were labeled with face position, gender, and skin tone information. The images were run through SaaS facial recognition platforms from Megvii Face++, IBM, and Microsoft. In all three of these platforms, the classifiers performed best on male faces (with error rates on female faces being 8.1% to 20.6% higher than error rates on male faces), and they performed worst on dark female faces (with error rates ranging from 20.8% to 30.4%). The authors hypothesized that this discrepancy is due principally to Megvii, IBM, and Microsoft having more light males than dark females in their training data, i.e. dataset bias.

In January 2019, researchers Inioluwa Deborah Raji and Joy Buolamwini published a follow-up paper that ran the experiment again a year later, on the latest versions same three SaaS facial recognition platforms, plus two additional platforms: Kairos, and Amazon Rekognition. While the systems' overall error-rates improved over the previous year, all five of the algorithms again demonstrated better than average error-rates for male faces, and significantly worse than average error-rates for dark female faces.

Gender classification of transgender and nonbinary individuals
Rekognition's gender identification technology categorizes faces as only male or female, with no other options. Critics have identified a number of disadvantages of this approach. First, there is no category for individuals nonbinary gender. Second, in an experiment conducted by a journalist, it was found that Rekognition is less accurate at identifying the gender of transgender individuals than of cisgender individuals.