La División Académica de Ingeniería invita a la Plática "Artistic Image Recovering from Principal Components", que impartirá: Jorge Isaac Chang Ortega, exalumno de la ingeniería en computación.

Abstract:

This work presents a comparison of different deep learning models for the reconstruction of artistic images from compact representations generated using Principal Component Analysis. The reconstruction models correspond to different types of Convolutional Neural Networks. Our results show that the statistics captured by the principal components transformation are enough to obtain good approximations in the reconstruction process, especially in terms of color and object visual features, even when using compact representations whose length is only about 1% of the original image space's total number of features.

Join Zoom Meeting

https://itam.zoom.us/j/94020421076?pwd=OFVjdjYxUWZya1dHVk1KR0Q4Z2hXUT09 

Meeting ID: 940 2042 1076