An extreme learning machine (ELM) is a feed-forward neural network (FFN). It is playing a crucial role in image classification and it is a simple method for features learning and ELM is frequently used in sequential learning, batch learning, and incremental learning because of these leanings ELM is learning faster, speed, good generalization ability, ease of implementation and efficiency in classifications. In this work, we propose two stages: the feature mapping stage, feature learning stage in an extreme learning machine.