RBE Practicum Presentation: Sameer Malik | Color Gamut Analysis for Deep Learning Trained Models

-

RBE Practicum Presentation

Sameer Malik

Color Gamut Analysis for Deep Learning Trained Models

Thursday, December 15th

1:00 PM - 2:00 PM

Zoom Link: Email smalik@wpi.edu

 

Abstract: In recent times, there has been a growing trend for involving deep learning models in critical decision making.But deep learning based models works like black boxes. It is often very hard to determine the parameters for which the models are well defined. Most of the time we don’t know how all the individual neurons work together to arrive at the final output. This become critical for applications where human safety is involved. Motional® is an autonomous car company which utilizes deep learning models for detection of traffic lights states. This presentation revolves around the traffic lights color gamut. Further it discusses the Color Gamut for which the model is trained and how to utilize that gamut to improve the detection results for traffic lights.

Practicum​ Advisors:

Professor Mohammad Mahdi Agheli Hajiabadi, Worcester Polytechnic Institute (WPI) 

Professor Berk Calli, Worcester Polytechnic Institute (WPI) 

Eric Luo, Senior Engineer, Motional

Audience(s)
DEPARTMENT(S): Robotics Engineering