This project aimed to carry out experiments, to find the "best" model to classify the different types of rock
For this project we used the dataset called "Rock Image" made available by Alexis Pascual on the website Mendeley Data
For this project, several models of CNN's were developed in order to find the best model. These tests consist of:
-
Create models that classify only 5 rock types at a time, increasing the amount of convolution layers of each new experiment
-
models that classify the 9 rock types at a time, increasing the amount of convolution layer creation each time new experiment
This project was carried out at the Nu[tec]² laboratory, available from IFES, under the guidance of Prof.ª Dra. Karin Satie Komati