I have combined the previous research on Fisher Vectors and Residual Enhanced Visual Vector to generate compact indices for database images to achieve efficient slide-to-lecture-video search.
In order to exploit the temporal redundancy of videos, an augmented database of clean slide images is used instead of the database of video keyframes.
Using augmented database has led to 100X memory reduction and has boosted the retrieval precision from 18% to 90%. An Android Application is implemented and tested on NVIDIA SHIELD tablet.
Ardunio code for Mobile Robot Localization based on Kalman Filter by MCECS-bot
There are no reviews yet.