Key takeaways from IMVC 2019, the Israel Machine Vision Conference
This year marks the 10th anniversary of the IMVC (Israel Machine Vision Conference), the largest conference of its kind in Israel. The conference brings together innovative startups and powerful companies to discuss the latest innovations in image processing, computer vision, machine learning, and deep learning. It also showcases research through student demos and posters.
Summing up all the amazing and futuristic innovations showcased at IMVC and its numerous conferences would be a challenge. Therefore, we chose to mention the ones we liked the most, especially around Augmented Reality. Augmented Reality is no longer a thing of the future, but is happening right now as we saw at IMVC.
Image Processing And Facial Recognition Will Become A Must
KanduAI presented its solution in the “Startups and Demos Zone” sponsored by BOSCH. Displayed was KanduAI’s facial and object recognition app. The application can capture an image from its surroundings or from any screen, detect and analyze people or objects in a fraction of a second.
Automotive AI : Simulate Real Life Situations to Anticipate Accidents
1 Billion. This is the number of miles one should drive in order to experience all possible hazardous scenarios. So, it’s no surprise that autonomous vehicles, expected to hit the market a few decades from now, are still delayed by the critical concern for safety. Some startups began paving the way for truly safe, autonomous cars. These companies use Artificial Intelligence R&D by training machines to understand traffic and anticipate all possible situations.
Guy Tsafrir, from Cognata Ltd., presented the company’s simulation platform, which provides a virtual road environment based on real maps. The software accurately simulates real-world test driving in order to support innovation in the field of autonomous vehicles. The Cognata platform uses deep learning technology to generate a 3D map of any city, including all roads, buildings and even traffic lights, out of photographic maps like those found on Google Earth®. The system then integrates historic data on local weather and lighting conditions. The result: a virtual reality simulator that can simulate real-life interaction between car sensors and actual driving situations. How does it work? Cognata’s technology is feeding in different times of day and driver behavior, in a lifelike environment of an actual city. It looks really cool, too! According to Tsafrir, Cognata’s deep learning platform can train ADAS systems to understand the billions of scenarios that may occur on the road.
Another exciting Israeli company, Vayavision, developed a cutting-edge 3D environmental model -this time for vehicle perception in the real world. Their model applies vision algorithms and deep learning to fuse data from their lidar system, which measures distances via cameras and radars, creating an upsampled high-resolution 3D model. This allows for high detection of all surroundings and potential threats facing autonomous cars. Vayavision’s system can detect objects and dangers even when sensors don’t work properly.
Self-driving cars may be all the rage these days, but railway transport is here to stay and train collisions remain all too frequent. RailVision, a start-up from Ra’anana, presented an Advanced Early Warning System that can automatically detect obstacles on a train’s path. They use advanced visualization and image processing to anticipate objects up to several kilometers away, in all weather conditions. The platform goes beyond visualization by monitoring infrastructure conditions in real time to predict malfunctions and by using Big Data to model behavioral and environmental trends that could put trains at risk of accidents.
Retail Is Being Digitized thanks to AI
How does the location of a product on a store shelf affect sales? Dolev Pomeranz presented
Trax Retail, a solution that can generate maps enriched with data, such as product category and location. Mobile robots scan all the areas of a store from all angles. Data is gathered and analyzed to understand how consumers react to product location and optimize how goods are laid out in retail areas. When entering in the store, customers can connect to an app and preselect their favorite products or discounted goods. The app displays a real-time map, guiding them to the right shelf.
The impact of artifical intelligence within every industry is truly fascinating as companies continuously look for challenging ways to improve and compete with each other.
To know more about KanduAI’s technology, please contact us.