Deep Learning Based Custom Image Multi-Class Classifier
Year: 2024
Type: dissertation
Abstract: This report details the design, and implementation of the multi-class image classifier using deep learning approach in Python with the help of TensorFlow. The data for this study were obtained from Kaggle and pre-processed and augmented through some transformations to make the final models more resilient. The model was trained using early stopping to avoid overfitting and the final model displayed a validation accuracy of 100%. In addition, the confusion matrix and training history plots show the remarkable accuracy of the model under consideration. This project demonstrates how the use of transfer learning and data augmentation enhances the performance of a model for the task of image classification.
Author: Abdullah, Mohammed M.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949814903
Cites:
Cited by:
Related to: