Breast cancer is the most common type of cancer worldwide. Early diagnosis of breast cancer can result in better treatment options increasing the survival chances of a patient. Automated or computer aided detection of breast cancer is applied in order to improve the accuracy and turnover time. However, the accuracy of automated detection systems can still be improved. Most of the efforts in the computer aided detection systems classify the images into cancerous and non-cancerous categories. The aim of this paper is to classify the mammographic lesions into three categories namely circumscribed, speculation and normal. The novel Pixel N-gram features have been used for classification of these lesions. Pixel N-grams are originated from character Ngram concept of text categorization. Classification performance is noted in order to analyse the effect of increasing N and effect of using different classifiers (MLP, SVM and KNN). It was observed that the classification performance increases with increase in N and then starts decreasing again. Moreover, classification performance achieved using MLP classifier was better than the performance using SVM or KNN classifiers. Keywords—Classification, Mammograms, N-grams, SVM, MLP, KNN