How Computer Vision is Changing the Vision of the Future
In December 2016, Amazon opened its first partially automated convenience store, AmazonGO, for its employees. In two years, AmazonGO opened its doors to the public.
Today, AmazonGO equips other retailers to use its technology, Just Walk Out (a combination of technologies that includes computer vision, sensor fusion, and deep learning), in their stores.
Just Walk Out, true to its name, allows the shopper to just walk out of a store, taking whatever, they want with them. The shoppers neither have to stand in line nor have to pay at the store. It may sound like theft, except it is not. AmazonGO’s Just Walk Out, is better at preventing theft than a conventional store.
While Just Walk Out is built for the customer’s ultimate shipping experience and convenience, it is also built to ensure the shop owner/retailer’s ultimate peace of mind.
However, the no-checkout convenience store powered by super surveillance systems does not in any way offer a 100% guaranteed effectiveness against shoplifting, at least not the unintentional shoplifting.
Despite the weight sensors installed in the shelves, AmazonGO stores might, from time to time, accidentally miss adding the item to the shopper’s cart if the shopper happens to take multiple yogurts at the same time. That said, every convenience story is designed to expect and accept a certain amount of lossage
The US Army has a vision for Computer Vision (CV).
It wants CV to acquire, to identity, and to engage targets directly – without any inputs from a human operator.
Currently, CV is only intended to be employed as a virtual solider – designed to aid a human analyst to take informed decisions based on the vast amounts of data CV gathers using high-quality satellite imagery.
Other than being able to deliver actionable information, CV cannot actually pull the trigger.
At least that is how the US Army’s ATLAS (Advanced Targeting and Lethality Automated System) program perceives CV’s scope.
But can Computer Vision pull the trigger on its own?
Yes, it can.
And that day is not too far into the future.
Computer Vision (with the help of AI and ML) can be deployed to do more than just aid humans in the analysis of vast amounts of data.
But there is resistance for the future.
Stopkillerrobots.org, the campaign to prevent the development and deployment of autonomous warfare equipment, has been against the push for unmanned/autonomous weapons that can identify and engage targets independently.
This has not stopped the Army Contracting Command (ACC) from asking military contractors, vendors, and academics specializing in computer vision to submit proposals to develop practical modules and integrations for the ATLAS (Advanced Targeting and Lethality Automated System) program.
According to the US Army, ATLAS is only intended to detect potential targets the humans might have missed.
Computer Vision has tremendous uses in healthcare.
A patient’s data can have up to 250GB of data in images and records. Being able to identify, evaluate, and interpret visuals and information like a human eye, computer vision helps doctors process an overwhelming amount of patient’s data in a short amount of time.
Microsoft’s InnerEye, a CV-oriented AI uses the latest Convolutional Neural Networks and voxel-wise segmentation of medical images to democratize AI for medical imaging – the most significant source of data in healthcare.
This process can accelerate clinicians’ ability to perform radiotherapy planning 13 times faster.
When doctors can spend less time in analyses, they can spend more time with patients.
For agriculture, soil and water management are essential. Farmers are forced to look for alternatives to meet growing demands. They have also had to mimic soil and water conditions for plants to grow. This is a complicated process. That is where computer vision comes in.
SWIR Vision Systems, a US-based startup, develops high-definition cameras capable of delivering wavelengths between 1.400 and 3.000 nanometers. For a perspective, the human eye can detect wavelengths from 380 to 700 nanometers. SWIR cameras can supply vital information such as water stress and moisture measurement, helping farmers use their resources efficiently.
By 2024, the CV market will grow at a CAGR of 7.8% ($17.4 Bn).
CV is gaining steady and exponential ground. Device manufacturers, part manufacturers, semiconductor companies, and software developers will be investing heavily in the infrastructure and development of CV.
Startups and prominent companies will develop prototypes that will get into full-scale production, depending on their success and need.
As the end consumers stand to be the ultimate beneficiaries of CV, a little publicity and knowledge-based information about CV can help them understand what CV is and how it can transform how they do things.
The talent landscape is rapidly changing. This is challenging HR teams to understand the evolving demands for new skills. Just as computer vision helps identify and prioritize targets that humans might have missed, Draup uses its proprietary AI/Big Data frameworks to help identify professionals with extremely specific skills that are needed for the CV future.