Robust Gesture Recognition Using Gaussian Distribution for Features Fitting

Abstract

Hand gesture became second language that complements almost many speeches, encounters, lectures, as well as friends chatting even in computer chatting they may gesturing to each other since it is a rotted habit in our behavior, we can notice that even if someone was sitting alone and thinking; he will continue gesturing during his meditation, however, the imitating of this natural behavior is an important issue for transferring this behavior to the human made machines and the intuitive interface will not be changed as compared to human-human communication, in this paper, we have applied a novel approach for recognizing the hand gesture and to maximize the level of unrestricted communication by solving the problem of rotation invariance that matters, we have employed a Gaussian bivariate likelihood function for hand modeling and features fitting and to produce uniform features that can be a reference for gesture database, we have achieved remarkable recognition percentages using 20 different gestures with a high speed recognition time, our system can be used for real time applications in which the time factor is important issue for the success of such systems, we have made a comparative study with some other known gesture algorithms as well.

Hand Gesture Recognition for Sign Language: A New Higher Order Fuzzy HMM Approach

“…The aim of the gesture recognition is to enable humans to interact with the human made machines in natural way without any mechanical devices and the mathematical equations will be the translator that translates the poses between the gestures and the telerobotic [1]. The gesture recognition is very difficult and complex task since the full recognition system should be able to identify the hand in different scales, positions, orientations, contrasts, luminosity, and others.…”

Hand Gesture Recognition for Sign Language: A New Higher Order Fuzzy HMM ApproachHand Gesture Recognition for Sign Language: A New Higher Order Fuzzy HMM Approach

“…In general, we can totalize the basic operations for gesture recognition system into four main operations; collecting the inputs, image preprocessing, feature vector extracting and efficient classification algorithm [1]. In the first stage, the meaningful gestures that will be used as an input for the gesture recognition system must be predisposed and decided.…”

“…Also the most important thing in hand gesture recognition system is the input features and the selection of good features representation [3]. There are some challenges that are facing the process of gesture recognition and these challenges can be found in preprocessing step and these are common for almost all the gesture recognition application [1], and if the gesture is not well obtained then this will affect all the latter pending process. The challenges can be summarized by the Fig.…”

“…Gestures are classified into two type static gestures and dynamic gestures. Static gestures refer to certain pattern of hand and finger orientation whereas dynamic gestures involve different movement and orientation of hands and face expressions largely used to recognize continuous stream of sentences [1].…”

“…To further increase the recognition rate and make the recognition system resilient to view-point variations, the concept of derived features (prominent features are derived from the available data set of features using forward selection algorithm) is introduced Hand Gesture Recognition for Sign Language: A New Higher Order Fuzzy HMM Approach Saad M. Darwish, Magda M. Madbouly, and Murad B. Khorsheed [4]. Readers looking for more information can refer to [1]. Gesture recognition schemes can be broadly classified into two groups.…”