scabini

RADAM

We propose a new method named Random encoding of Aggregated Deep Activation Maps (RADAM) for feature extraction from pre-trained Deep CNNs. The technique consists of encoding the output at different depths of the CNN using a Randomized Autoencoder, producing a single image descriptor

MLP-CN_dataset

A dataset of fully-connected neural networks trained on vision benchmarks. Data contains all their synapse states, i.e., from random initialization and throughout all training epochs. Complex Network measures are then computed to each of their hidden neurons.

SSN

Code for Spatio-Spectral Networks (SSN), a method to compute color-texture descriptors from images

PArewiring_weights

Weight Organization Matters: Improving Deep Neural Network Random Initialization Through Neuronal Rewiring

ADCN

Angular Descriptors of Complex Networks for Shape Analysis

COmplexVID-19

Simulate COVID-19 epidemic spreading with Complex Networks, given demographic data of a given community, and control restriction levels on different social interaction layers.

covid_genosensors_ML

keras-yolo3

Training and Detecting Objects with YOLO3

mechanosensors

Code for the Image processing and Machine Learning techniques employed for the mecanosensor recognition

mypy

Optional static typing for Python

stable-diffusion-evolved

A latent text-to-image diffusion model