Image classification is a critical domain in computer vision with diverse practical applications. Recent advancements in deep learning have shown significant success in this field, with the COREL 1K dataset serving as a standard benchmark for evaluating such algorithms. This paper delves into an extensive investigation of machine learning (ML) based image classification using the COREL 1K dataset. The study comprehensively explores various ML algorithms, including decision trees, k-Nearest Neighbors, and Support Vector Machines (SVM), comparing their performance on the dataset. Additionally, the research scrutinizes the impact of different feature extraction techniques, such as Local Binary Patterns (LBP), color histograms, and Gray-Level Co-occurrence Matrix, on the accuracy of classification. The experimental results on the test set reveal noteworthy insights. Specifically, SVM employing the LBP feature extraction method achieves the highest accuracy at 87.5%. This outcome underscores the efficacy of both feature extraction methods and ML algorithms in the context of image classification on the COREL 1K dataset. By systematically evaluating and comparing these approaches, the study contributes valuable insights into optimizing image classification techniques, offering a nuanced understanding of the interplay between ML algorithms and feature extraction methods for enhanced accuracy in real-world applications.