Amazon's AI has the ability to guess your age
Not only can Amazon predict what products you're most likely to purchase, but it is also has the ability to guess your age.
The firm's AI, Rekognition, received an update that provides an estimated age range of a person in an uploaded image – the value is expressed in years and is returned as a pair of integers.
Amazon believes this new attribute can be used to power public safety applications, collect demographics or create a timelapse in photos.
Jeff Barr, chief evangelists at Amazon Web Services, explained in a recent blog post that the AI offers a 'fairly wide range' when it makes its guess about a person's age.
'In general, Rekognition the actual age for each face will fall somewhere within the indicated range, but you should count on it falling precisely in the middle,' he wrote.
President Donald Trump is 70 years old and the AI guess a range from 60 to 90 years old, which coincides with Barr's explanation.
However, the AI was somewhat off when it came to suggesting the age group for Jennifer Lopez – she is 47 years old and Rekognition guessed 26 to 44 years old.
Amazon Rekognition is a developer toolkit that is part of the firm's AWS cloud computing service.
Using the technology, applications will have the power to detect objects, scenes and faces within images.
And it was built by computer vision and uses deep learning neural network models to analyze billions of images daily.
Barr said the feature is now available and can be demoed by anyone.
'The APIs find and compare faces, detect thousands of objects,' Amazon AWS explained.
'We continue to add new objects and make improvements to facial analysis so you can focus on building applications'
The technology has the ability to comprehend scenes, objects and faces.
'Given an image, it will return a list of labels. Given an image with one or more faces, it will return bounding boxes for each face, along with attributes,' explained Barr.
He used an image of his golden retriever, Luna, as an example of how well the AI can label objects, people and animals.
'Rekognition labeled Luna as an animal, a dog, a pet, and as a golden retriever with a high degree of confidence,' explained Barr.
'It is important to note that these labels are independent, in the sense that the deep learning model does not explicitly understand the relationship between, for example, dogs and animals.'
'It just so happens that both of these labels were simultaneously present on the dog-centric training material presented to Rekognition.'
Barr explained that there is a lot of work that goes into training the AI to detect faces.
'In essence, you present the learning network with a broad spectrum of labeled examples ('this is a dog', 'this is a pet', and so forth) so that it can correlate features in the image with the labels,' he noted.
And the phase is also 'computationally expensive due to the size and the multi-layered nature of the neural networks'.
'After the training phase is complete, evaluating new images against the trained network is far easier,' explained Barr.
'The results are traditionally expressed in confidence levels (0 to 100%) rather than as cold, hard facts.'
'This allows you to decide just how much precision is appropriate for your applications.'
Barr also shared another example with him and his wife.
Rekognition was able to find their faces, setup bounding boxes around them and let him know if he and his wife were happy in the picture, if they were wearing sunglasses if either of them had a beard and more.
The AI can be used in a range of applications - if you have a large collection of photos you c an tag and index them using the technology.
And users have the ability to 'process millions of photos per day without having to worry about setting up, running, or scaling any infrastructure'.
'You can implement visual search, tag-based browsing, and all sorts of interactive discovery models,' Barr shared.
'You can compare a face on a webcam to a badge photo before allowing an employee to enter a secure zone.'
'You can perform visual surveillance, inspecting photos for objects or people of interest or concern.'
'You can build 'smart' marketing billboards that collect demographic data about viewers.'
MTCHT
ICT
TECHNOLOGICAL INNOVATIONS
POST
ABOUT US
NEWS
INTERESTING
INTERVIEW
ANALYSIS
ONLAIN LESSONS
