They evaluated this algorithm on human action detection using a public video with cluttered background and detection of customers showing an intention of purchasing the items on the shelf in a shopping mall. (A study on human action analysis that also supports the summative introduction; a study that proposes and eventually led to the creation of the of an algorithm for detecting human actions called SMILE-SVM) Moreover, in 2009 Ming-Yu Chen and Alexander Hauptmann had a study that led to the foundation of MoSIFT, an algorithm that detects interest points to analyze human motion. It is based on the SIFT, a well-known descriptor. It resulted on a 95.8% accuracy on KTH dataset, a database of human actions like walking, jogging, running and more. (A study that also support the summative introduction; a study that also presented another algorithm for detection of human