Comparative analysis of image classification algorithms based on traditional machine learning and deep learning
Manoj Burman1,Ranjan Singh2,Rupal Das3,Srishti Gauraha4,Sanit Kumar5 , Prof. Prince Sahu6.
1B.Tech 8th Sem Student, Computer Science and Engineering, Government Engineering College, Bilaspur, Chhattisgarh, India
2B.Tech 8th Sem Student, Computer Science and Engineering, Government Engineering College, Bilaspur, Chhattisgarh, India
3B.Tech 8th Sem Student, Computer Science and Engineering, Government Engineering College, Bilaspur, Chhattisgarh, India
4B.Tech 8th Sem Student, Computer Science and Engineering, Government Engineering College, Bilaspur, Chhattisgarh, India
5Asst. Professor, Department of Computer Science and Engineering, Government Engineering College, Bilaspur, Chhattisgarh, India
------------------------------------------------------***----------------------------------------------------------
Abstract
Image classification is a critical task in computer vision that has applications in various domains. Traditional machine learning algorithms and deep learning techniques are commonly used for image classification. This research paper presents a comprehensive comparative analysis of these two approaches. The objective is to evaluate and compare the performance, computational complexity, interpretability, and robustness of traditional machine learning algorithms, including SVM, random forests, and KNN, with deep learning algorithms, primarily focusing on CNNs. The analysis is conducted using a specific image classification dataset, and performance evaluation metrics such as accuracy, precision, recall, and F1 score are considered. The findings provide valuable insights into the strengths and limitations of each approach, enabling researchers and practitioners to make informed decisions when selecting image classification algorithms. The study also highlights potential future research directions in the field of image classification..
Keywords: Image classification, Deep learning, Convolutional Neural Network, CIFAR-10, Optimization, ImageNet, Dimensionality Reduction, K-Nearest Neighbours.