Convolutional Neural Network (CNN) responses when the number of classes increase

From ISLAB/CAISR
Title Convolutional Neural Network (CNN) responses when the number of classes increase
Summary Convolutional Neural Network (CNN) features behaviour in the context of textures
Keywords
TimeFrame
References
Prerequisites image analysis
Author
Supervisor Josef Bigun, Fernando Alonso-Fernandez
Level Master
Status Internal Draft

Generate PDF template

Convolutional Neural Network (CNN) features behaviour in the context of textures

The project aims to quantify the behaviour of Convolutional Neural Network (CNN) features in the context of textures.

Image analysis

Josef Bigun, Kevin Hernandez-Diaz, Fernando Alonso-Fernandez.

The project aims to quantify the behaviour of Convolutional Neural Network (CNN) features in the context of textures.

Master, Open

First it will Investigate how the cnn's behave when the number of classes increase, potentially to infinity. Here, syntethic images can be used since the groundtruth is known. One such classificaiton problem comprises finding the local orientations on a series of Frequency Modulated test images and associated certainties. The number of classes can be made as large as one wishes in both cases, the discretizaton grid being the limit (In FM-test images this equals to their sizes).

Second, Convolutional Neural Network (CNN) features discrimination power will be be quantized in terms of within and between classsvariances, with different textures in a mosaic of texture images, cut from real aerial images and Brodatz textures.