Data Scientist at EPAM
Irakli Salia is a Data Scientist at EPAM with interest in machine learning and optimizing the models for edge deployment. He specializes in computer vision and generative AI and is currently actively pursuing an MA in computer science at Ilia State University while holding a BA in Business Administration from Free University of Tbilisi.
Workshop: Quadrilateral object detection in retail scene and enforcing temporal consistency in Computer Vision: Data Efficient approach
In this talk, I will present the process we followed to build a quadrilateral object detection model for the retail scene and how we have deployed this model to mobile devices. I will also share how we managed to increase the temporal consistency in our application without any additional data collection and without adding any extra parameters to our original model.