A Neural-Network-Based Approach to Recognizing 3d Arm Movements

Abstract

Gesture recognition is needed for a variety of applications. One particular application of gesture-based systems is to implementIn general, hand gestures can be described by the computes has led to considerable interest in recognizfollowing four attributes: (1) hand shape (i.e. posture); ing hand gestures. For example, it is a more intuitive (2) palm orientation; (3) hand location; and (4) arm and natural way to use a so-called finger mouse that movement. The attributes can broadly be separated enables a user to specify commands and additional painto static (spatial) and dynamic (spatio-temporal) atrameters by drawing single intuitive gestures with his tributes. Generally, the values of the static attributes or her finger. A variety of gesture-based applications will be calculated directly from the data supplied by have been created so far. The use of hand gestures as the hardware, whereas the dynamic attribute values input interfaces is widely addressed in several articles will be calculated from the values of a sequence of (to name just a few here) [1]- [15].static attribute values. Since hand gestures are spatioOne particularly important application of gesturetemporally varying, the automatic gesture recognition based applications is to implement speaking aids for turns out to be very challenging. Many different apthe deaf [16]- [23]. Fig. 1

A rule extraction based approach in predicting derivative use for financial risk hedging by construction companies

“…Halqamuqe and Glesner proposed an architecture for neurofuzzy based rule extraction in 1992 (Halqamuqe & Glesner, 1992). Su developed HRCNNs to accelerate the learning process, and later, successfully implemented the method for recognizing arm movements and medical diagnosis (Su, 1993;Su, Zhao, & Lai, 2003). It is believed that the rule extraction based approach is an effective method for problems related to pattern classification (Su et al, 2003;Tsai, Cheng, & Chang, 2006;Wong, Lin, & Chen, 2005).…”

“…Su developed HRCNNs to accelerate the learning process, and later, successfully implemented the method for recognizing arm movements and medical diagnosis (Su, 1993;Su, Zhao, & Lai, 2003). It is believed that the rule extraction based approach is an effective method for problems related to pattern classification (Su et al, 2003;Tsai, Cheng, & Chang, 2006;Wong, Lin, & Chen, 2005). Since the problem dealt with in this study is a typical pattern classification prediction problem, the HRCNNs approach is chosen as the rule extraction based approach and its establishment is introduced in the following paragraphs.The HRCNNs essentially integrate neural networks using rule extraction, rendering a tool for machine learning and for pattern classification.…”

“…1 indicates a two-layer HRCNN for this model. Given that the output is defined as OutðxÞ, the output function can be expressed as (Su et al, 2003) OutðxÞ…”

Tendencias investigativas en el estudio de tecnologías inclusivas para población sorda

“…En cuanto al tercer término más utilizado, como es el caso del reconocimiento de gestos (Gesture recognition), se relacionan distintas exploraciones acerca de la búsqueda de un lenguaje común entre personas sordas, mediante el uso de dispositivos móviles inteligentes, empleados en sistemas de interpretación de lenguaje de señas a través de sensores Fusión (Lee & Lee, 2018); adicionalmente, son diversas las revisiones asociadas con el uso de tecnologías en 3D, no invasivas y programables, para el reconocimiento de gestos; o bien, las indagaciones apuntan hacía diferentes perspectivas, con el fin de reconocer los movimientos en 3D, basados en redes neuronales o enfoques difusos que faciliten el proceso de comunicación en tiempo real, de modo interactivo y personalizado; siendo de utilidad (habilitadores de la interacción entre el hombre y la máquina) el uso de algoritmos que permiten el reconocimiento de signos y rastreo manual en 3D, integrado en cámaras con funcionalidades especiales (estéreo binocular), para la recopilación de información importante de la profundidad que presente un objetivo (Paudyal et al, 2016;Su et al, 2001;Su et al, 2003;Yang et al, 2017).…”