PyData Tel Aviv 2022

Synthetic Data in Vision DL for higher accuracy, faster training
12-13, 11:15–11:45 (Asia/Jerusalem), Track 1

Synthethic Data Generation is now a popular and important addition to standard image augmentation methods and real life data acquisition and labeling. We will show how the unique, metadata-driven nature of synthetic data pipelines enables new method to improve AI explainability, accuracy and training times


NEC has been using Synthetic Data for Vision based DL heavily for both 'people in the wild' situations and applications (detection, pose estimation) and for medical imagery AI development. In the talk we show how rather than 'spray and pray' large synthetic datasets, unique, precision guided generation, iterative with the DL models parameters, can yield much faster training, overall accuracy and provide novel explainability mechanisms.


Prior Knowledge Expected

Previous knowledge expected

Tsvi Lev is the Managing Direcor of NEC Corporation's Israeli Research Center, the first Open Innovation center of NEC globally. The center is engaged in applications of AI to Cyber, Physical Safety and Medical.

Before that Tsvi was VP of Strategy at Amdocs, leading the Amdocs team in the AT&T Open Innovation Organization – the AT&T Foundry.

Previously, Tsvi managed the R&D Center of Samsung Electronics in Israel, a center devoted to innovation and cutting-edge development for Samsung’s future consumer electronics and cloud products.

Tsvi was an Entrepreneur in the Mobile Multimedia and Computer Vision space, and sold his company to Emblaze Systems.

A Talpiot Graduate, Tsvi holds a Masters in Theoretical Physics, is the author of numerous patents, and a frequent speaker at events related to technological innovation.

Or Katz