Kepler Vision Technologies will be present at the LDV Vision Summit in New York. We would love to talk to you! Fill out the contact form here, so we can catch up. Other events where we will attend are the following:
- June 2018 at the CVPRConference in Salt Lake City, USA;
- September 2018 at the ECCVConference in Munich, Germany.
Trends in investing in visual technologies
The LDV Vision Summit hosts panels on trends in investing in visual communication and visual technologies. This makes it a unique conference to go to. A few of last year’s discussions are the following:
- Cameras creeping into our lives,
- Applications for those who are visually impaired,
- How people navigate inside a store,
- Inspections of power lines using drones.
Three keynotes on using synthetic data
The other main attractions of the 2018 LDV Vision Summit are these three keynotes on using synthetic data to train neural networks:
- Retail Brain Using All Synthetic Data And Going After Amazon Go. Acquiring high-quality data with accurate labels is both time-consuming and expensive. Often, it may not even be possible. Ying Zheng (Chief Science Officer from AiFi) will share how AiFi leverages synthetic data to create checkout-free solutions at massive scale. AiFi creates large-scale store simulations and renders high-quality images with pixel-perfect labels. These are then used to train deep learning models for camera calibration, multi-camera multi-people tracking, re-identification, and product recognition.
- Synthetic data will help computer vision make the jump to the real world. Modern deep learning and artificial intelligence are data-hungry. Josh Tobin (Research Scientist at OpenAI) will share how computer vision will bridge the gap in applications where data collection is expensive. He will make the case for synthetic data, describing how OpenAI uses this to learn perception models for real-world robotic control.
- Synthetic Data for AI Model Development Neuromation aims to democratize artificial intelligence. It does so through the use of synthetic data and distributed computing to dramatically reduce the cost of development. Neuromation’s CEO Yashar Behzadi explains how they built a “token-based” global marketplace. Their platform connects AI talent, data providers, and customers to enable the development of novel AI solutions.