In this thesis, a dynamic gesture recognition system is presented which requires no special hardware other than a Web cam . The system is based on a novel method combining Principal Component Analysis (PCA) with hierarchical m ulti-scale theory and Discrete Hidden Markov Models (DHMMs). We use a hierarchical decision tree based on multi-scale theory. Firstly we convolve all members of the training data with a Gaussian kernel, w h ich blu rs d iffe ren c e s b e tw e en images and reduces their separation in feature space. Th is reduces the number of eigen vectors needed to describe the data. A principal component space is computed from the convolved data. We divide the data in this space in to several clusters using the £-means algorithm. Then the level of b lurring is reduced and PCA is applied to each of the clusters separately. A new principal component space is formed from each cluster. Each of these spaces is then divided in to clusters and the process is repeated. We thus produce a tree of principal component spaces where each level of the tree represents a different degree of blurring. The search time is then proportional to the depth of the tree, which makes it possible to search hundreds of gestures with very little computational cost. The output of the decision tree is then input in to the DHMM recogniser to recognise temporal information.