Multi-HMM classification for hand gesture recognition using two differing modality sensors

Abstract

This paper presents a multi-Hidden Markov Model (HMM) classification approach for hand gesture recognition by utilizing two differing modality and low-cost sensors. The sensors consist of a Kinect depth camera and a wearable inertial sensor. It is shown that the multi-HMM classification based on nine signals that are simultaneously captured by these two sensors leads to a more robust recognition compared to the situation when only a single HMM classification is used to generate the likelihood probabilities of hand gestures. This approach is applied to the hand gestures of the $1Unistroke Recognizer application and the results obtained indicate a 7% improvement in the overall classification rate over a single HMM classification under realistic conditions.

DEICTIC: A compositional and declarative gesture description based on hidden markov models

“…Besides the observation features, for training the HMM we need to define its number of states and the transition topology. Temporal processes like gestures and speech recognition are well suited for the left-to-right (or Bakis) topology [19,25], which includes arcs between two states s i and s j only if j ≥ i or, equivalently, if the transition probability matrix A is upper triangular. In this paper, we use a Bakis topology for the ground terms, but the composition works also if different topologies are used (e.g., ergodic) [24].…”

Wearable sensors for gesture analysis in smart healthcare applications

“…The results suggest that the smartwatch could measure the tremor with good correlation with respect to other devices. In other research, Liu et al [48] fuse the information of an inertial sensor placed on the wrist with a Kinect camera to identify different gestures (agree, question mark, shake hand) to improve the interaction with a companion robot. Ahanathapillai et al [10] use an Android smartwatch to recognize common daily activities such as walking and ascending/descending stairs.…”

The use of machine vision to control the basic functions of a CNC machine tool using gestures

“…A classification or recognition mechanism is a computational algorithm that takes the representation of an object and classifies it as some known class type. The most commonly used techniques for classification are the HMM (hidden Markov model) [30,31] and other statistical methods, ANN (artificial network neutron) [32], methods based on state space [33] wavelet transform [32], matching the curves [34], and database methods [35]. Since the issue of recognition of gestures is currently the subject of intensive research, there are many review papers in which authors present their research in detail and compare algorithms for segmentation, extraction, tracking and classification systems used in gesture recognition.…”