CNN-for-Predicting-Attachment-Level-of-Bacteria

This repository implements a Convolutional Neural Network (CNN) for classifying and predicting biomaterial attachment levels. It supports both regression and classification tasks, enabling precise analysis of biomaterial interactions. The CNN model is designed to handle complex data, providing accurate predictions and classifications to advance research in biomaterial science.

This Highly scalable framework allows for easy addition of more combinations in the future and can be seamlessly transferred to other projects.

Project Structure

• models/:

  • CNN.py: Implementation of CNN Architecture for Supervised-Learning.
  • Networks.py: Implementation of Base Network (parent) definitions and configurations for the CNN architecture.

• options/:

  • base_options.py: Basic Command-line arguments for the training script.
  • train_options.py: Hyperparameter Command-line arguments for the training script.

• utils/:

  • images_utils.py: Utilities for image handling.
  • visualizer.py: This file provides scripts for a TensorBoard visualizer for tracking training progress.
  • weights_init.py: This file contains scripts for weight initialization functions for the CNN architecture.

• train.py: Script for training the model.

• call_methods.py: This file contains scripts for dynamically creating models, networks, datasets, and data loaders based on provided names and options.

Requirements

To run the code, you need the following:

Python 3.8 or above PyTorch 1.7 or above torchvision tqdm matplotlib TensorboardX 2.7.0 Install the necessary packages using pip: