Classification of images with balanced class for faster retrieval of big data and accuracy using CNN

Hassan Mashraa AlBuqami

As huge amounts of structured and unstructured data are available in big data, faster retrieval of the needed information in a faster and more accurate way is essential. The classified data will be helpful to retrieve the big data in an effective way. CNN (Convolutional neural networks) is one such technique to deal with huge volumes of images. CNN’s are widely used in images and videos that extract the object’s features. To overcome the drawbacks related to class imbalance, a new clustered and centroid-based algorithm is suggested to balance the class. This helps in proper classification and to deal with applications where the minimum class data would play a major role e.g., Data of patients suffering from cancer. The proposed cluster-based classification method will help balance the class and retrieve the most accurate data faster. 

Advanced Studies: Euro-Tbilisi Mathematical Journal, Vol. 16,  supplement issue 2 (2023), pp. 57-65