Get more from AI Weekly
Breaking stories as they happen. Deep dives on the topics you care about. 50+ free courses from Stanford, MIT, and more.
Deep Dives Daily Alerts Learning AI
Embedding safety into in computer vision by improving data quality
According to the World Health Organization, there are 285 million visually impaired people around the world who not only live in darkness, but also suffer poverty caused by visual impairment. For those without the visual ability to perceive the outside world, AI can be used to convert the content or text they need to “see” the world so that they can still perceive the outside world with their other senses.
Hugo Tsai, Product VP of AngelEye’s company NextVPU, shared his team’s experience of using the Microsoft Trove saying for the AngelEye Smart Reader, an intelligent edge device based on computer vision that helps the vision impaired community, “We need a large number of pictures, and our internal team struggles to find the right quantity and breadth of categories. Trove connects us to partners all over the world who provide tremendously rich pictures”.
After using Trove to gather training data, Hugo Tsai found that the OCR character recognition accuracy has increased from 50% to 99.15%, which has significantly improved the experience of end users.
Sponsor
Get a $500 credit* towards image data collection costs using Trove
Trove is a new crowdsourcing marketplace from Microsoft where you can gather images for AI responsibly.
You get:
- Value: You don’t have to worry about whether you are getting the best value for photos obtained through Trove.
- Confidence: Data you collect will be done safely and responsibly through terms that respects the rights of submitters.
- Quality: Highly relevant and diverse images from real people, licensed for your specific scenario and use case.
*Eligibility for the $500 credit depends in part on having an acceptable Trove project and making payments via Trove. Full details on the offer and all eligibility requirements can be found in the official terms.
Find us on Facebook - Twitter - Linkedin
Research
The top 10 computer vision papers in 2020 with video demos, articles, code, and paper reference.
Here are my top 10 of the most interesting research papers of the year in computer vision, in case you missed any of them. In short, it is basically a curated list of the latest breakthroughs in AI and CV with a clear video explanation, link to a more in-depth article, and code
CLIP: Connecting Text and Images
We’re introducing a neural network called CLIP which efficiently learns visual concepts from natural language supervision.
iGibson Challenge 2021 - Interactive and Social Navigation in Indoor Environments
This year we present this challenge focused on the Interactive Navigation and Social Navigation problems, and invite researchers around the world to push the frontiers of vision-based indoor autonomous navigation by innovating with models and algorithms applied to these benchmarks.
How Much of a Photo Can Be Deleted Before AI Can’t Recognize It?
In a new project that mixes science and art, artistic duo Shinseungback Kimyonghun has created a series of images that have pixels removed until an AI program can no longer recognize the subject — in this case mountains.
Applied use cases
How AI vision systems can succeed with human input
Before trusting an AI model to make critical decisions one must subject it to robust testing and validation.
How an Automotive Tier 1 Supplier Used Atlas to Improve Computer Vision Accuracy By Up to 48% mAP Within Days
Computer vision is a fundamental technology in automotive Advanced Driver Assistance Systems for pedestrian, vehicle, or road sign detection. However, today’s vision architectures are usually designed and tuned manually for pleasing visual image quality .
AI and IoT – 5 use cases where it’s gathering pace
While IoT sensors detect external information, replacing it with a signal that humans and machines can distinguish, it’s AI that helps to build intelligent machines.
By bringing AI to surgeries, startup hopes to close knowledge gaps
Theator, based in Palo Alto with an R&D site in Tel Aviv, raises $15.5 million in a Series A funding round to grow its surgical intelligence platform
Market News
Image Recognition: Everything You Need to Know (Guide)
In this article we will cover image recognition, an application of Artificial Intelligence (AI) and computer vision.
Artificial intelligence: The dark matter of computer vision
What makes us humans so good at making sense of visual data? That’s a question that has preoccupied artificial intelligence and computer vision scientists for decades.
Ethics
The Capitol riot is spurring new interest in gun-detection AI
Facial detection has all sorts of technological and ethical problems. What if we tried to prevent crime by using software to spot weapons?
HUMANS ARE TRYING TO TAKE BIAS OUT OF FACIAL RECOGNITION PROGRAMS. IT’S NOT WORKING–YET.
“The way that we’re testing fairness in algorithms doesn’t really ensure fairness for all,” Khan says. “It may only ensure fairness in the vague, stereotypical sense of fairness.”
Healthcare
From Hobart, to London, to Dhaka: Using cameras and AI to build an automatic litter detection system
It’s estimated about two million tonnes of plastics enter the oceans from rivers each year. But our waterways aren’t just conveyor belts transporting land waste to the oceans: they also capture and retain litter.
AI Coupled With Videos of Movement May Help in Diagnosing Parkinson’s
This non-invasive, artificial intelligence technique has the potential to be a so-called second opinion to help physicians make an accurate Parkinson’s diagnosis more quickly, its researchers said.
AI uses "ugly duckling" technique to spot melanoma with high accuracy
A new deep learning algorithm can spot features of odd skin lesions and detect melanoma with high accuracy. Artificial intelligence is starting to combine with smartphone technology in ways that could have profound impacts on the way we monitor health
What do you think?
Join the conversation — share your take on this issue.
Log in to comment