Home Free Lab ReportsConclusion4


“In this paper we have attempted to present the theory of hidden Markov models from the simplest concepts (discrete Markov chains) to the most sophisticated models (variable duration, continuous density models). It has been our purpose to focus on physical explanations of the basic mathematics; hence we have avoided long, drawn out proofs and/or derivations of the key results, and concentrated primarily on trying to interpret the meaning of the math, and how it could be implemented in practice in real world systems. We have also attempted to illustrate some applications of the theory of HMMs to simple problems in speech recognition, and pointed out how the techniques could be (and have been) applied to more advanced speech recognition problems”

“In this paper, we attempt to recognize actions in real time on the Android platform, and recognize the user’s activities from a recognized set of actions. The two step HMM structure is appropriate for mobile environment to reduce computational complexity. Looking at the results of action recognition, there is confusion between ‘stair up/down’ and ‘walk’ actions. Careful data selection and training will improve the performance of recognition. Also, ‘shopping’ and ‘taking a bus’ activities have some difficult patterns to classify, and it is necessary to use more sensors such as a GPS receiver, Wi-Fi, etc”