Facial expression based automatic emotion detection is important in several research fields, health, security and human computer interfaces. Recently, researchers have shown keen interest for the feature extraction using emerging techniques in order to achieve accurate results. The main steps in emotion detection include the image preprocessing, feature extraction followed by the classification of these features to detect emotions. Facial landmarks refer to key points on a human face and represent those regions in a face which help to distinguish between emotions. The paper starts by discussing the importance of expressions and the role they play in emotions thus concluding expressions can serve as an important input to detect emotions. This paper presents a comparative analysis of different artificial intelligence techniques initiated for automated facial expression recognition and explains the process followed including the preprocessing techniques, the use of the right neural network for that approach (e.g. CNN and ANN). Moreover the impact of each study including the benefits and flaws of AI techniques have been discussed which will help further in improving the recognition system. The paper will focus on the approach taken by different emotion detection algorithms with different datasets and highlighting the key features.