Feature vector or time-series – comparison of gestures representations in automatic gesture recognition systems

Journal Title: Challenges of Modern Technology - Year 2015, Vol 6, Issue 1

Abstract

In this paper, we performed recognition of isolated sign language gestures - obtained from Australian Sign Language Database (AUSLAN) – using statistics to reduce dimensionality and neural networks to recognize patterns. We designated a set of 70 signal features to represent each gesture as a feature vector instead of a time series, used principal component analysis (PCA) and independent component analysis (ICA) to reduce dimensionality and indicate the features most relevant for gesture detection. To classify the vectors a feedforward neural network was used. The resulting accuracy of detection ranged between 61 to 87%.

Authors and Affiliations

Katarzyna Barczewska, Wioletta Wójtowicz, Tomasz Moszkowski

Keywords

Related Articles

What is at risk? – an environmental health risk assessment related with pharmaceutical substances in drinking water

Constantly increasing rate of drugs’ consumption can be related with environmental abundance of drug substances and their degradation products. Unused and used drugs are often discharged into municipal wastewaters or lan...

Color gamut volumes of ink-jet prints made on papers suitable for offset technique

Ink-jet as a digital technology can be used for carrying out very short runlenghts. This constitutes a very important issue nowadays since the digital printing market was estimated to grow by 71% from 2008 to 2013 year....

Modelling of different cogeneration technologies in a power generating unit with pwr reactor in the context of a future polish nuclear power station

This paper focuses on the analysis of technical possibilities of a nuclear combined heat and power (NCHP) unit. The possibility of implementing a cogeneration unit in the cooperation with a typical PWR (Pressurized Water...

Variability of roughness parameters calculated using data obtained by TLS depending on scan resolution and beam's angle of incidence

Spatial data obtained by Terrestrial Laser Scanning (TLS) can be used in order to create various inventories and analyses of the surveyed surfaces. This paper presents a use of orthogonal regression line in order to iden...

Integration of advanced oxidation process with nanofiltration for dairy effluent treatment

The paper presents the research results on the possibility of the integration of advanced oxidation process (AOP) involving Fenton reaction with nanofiltration (NF) for dairy effluent treatment. It has been found that Fe...

Download PDF file
  • EP ID EP206669
  • DOI -
  • Views 102
  • Downloads 0

How To Cite

Katarzyna Barczewska, Wioletta Wójtowicz, Tomasz Moszkowski (2015). Feature vector or time-series – comparison of gestures representations in automatic gesture recognition systems. Challenges of Modern Technology, 6(1), 33-37. https://europub.co.uk/articles/-A-206669