We present a framework for collecting large scale multimodal data sets useful for studying human attention mechanism through an eye tracking data analysis. Two approaches of data collection are presented. First - we have designed a low cost, acceptable in the public setting wearable hardware for recording point of view scene, audio, location and human eye movements. Second, we present a web based framework for crowdsourcing a human visual attention corpus that can be completed by annotating online video data with an eye tracking information. The power of our framework is that it is based entirely on low cost, commodity hardware, it is open source and does not require any software to be installed on the user side, thus allowing to use inexpensive labor pools available through crowdsourcing Internet marketplace and scale up the data set.
Dmitry Chichkov (NVIDIA)
Research Engineer, Machine Learning. 15+ years of experience in the industry. I'm currently with NVIDIA, Autonomous Driving, Perception,