Amin Kashi is a masters student in Autonomous Robotics team of ARAS group. His research interests include machine learning, computer vision, and robotics. His current project focuses on the perception of depth and semantics by DL methods. He loves testing his research results in 3D virtual environments (like PC games :).
There have been some researches showing relationships between the perception of depth and the perception of semantic segmentation. Although we intuitively accept that all visual perceptions correlate somewhat with each other, there should be a way to prove it more formally.
I’m currently working on depth estimation and semantic segmentation and their relationships in a single image. My goal is to improve the accuracy and performance of DNNs used for extracting these concepts from an image.
I’m also examining the effect of training a DNN on a synthesized dataset (from 3D modeling software and game engines) and then testing it on real data.
Please give me a message.