HEp-2 cell image classification is an important and relatively unexplored area of research. This paper presents an experimental analysis of five different categories of feature sets with four different classifiers to determine the best performing combination of features and classifiers. The analysis is performed on the ICIP 2013 Cell Image Classification Contest Training dataset comprising over 13,000 cell images pertaining to six cell classes. The results computed with 10 fold cross validation show that texture features perform the best among all the explored feature sets and the combination of Laws features with SVM yields the highest accuracy. © 2013 Springer International Publishing.