Enhanced performance of the different ML classifiers on Image datasets

Author : Varna C V

Abstract : Classification technique classifies the objects as two classes or multiple classes. Histogram is calculated for the dataset samples at the initial stage. Then SIFT algorithm is applied to the histogram processed dataset. After applying SIFT algorithm features are extracted from the image dataset and different Machine Learning models are experimented upon this processed dataset and its performance is compared. Then, Classifying features by computing distance between the points in the feature vector. If a point falls below the Threshold It is recognized as it belongs to class 0 otherwise to class 1.

Keywords : These improved classifiers result in more accurate and better performance than the traditional ones.

Conference Name : International Conference on Machine Learning Models and Applications (ICMLMA-26)

Conference Place : Bangalore, India

Conference Date : 10th May 2026