Neuronal Networks models used for the class of Machine Perception In this repository you will find the models implemented for the project 3 of Gaze Estimation of the class of Machine Perception at ETH Zurich at 2019.
We implemented 4:
-
A simple convolutional NN It was the simplest Net we used for training and to have an starting point to start. Table 1 shows the structure of this primitive Net. After that the signal was pass through a flatten layer. Then dense layers were used to obtain the output.
-
The AxelNet In a next step we decided to implement the AlexNet[1] model for the eye gaze estimation.
-
VGGNet In a second model we implemented a variant of the VGG Net as described in [2].
-
DenseNet
The rest of the framework, is distributed by the Assistants of the class and belongs to them.
References:
[1] http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf
[2] Karen Simonyan and Andrew Zisserman. 2015. Very Deep Convolutional Networks for Large-Scale Image Recognition. CoRR abs/1409.1556 (2015).