HAND SIGN RECOGNITION USING LOCAL FEATURE DESCRIPTOR WITH ELM CLASSIFIER

HAND SIGN RECOGNITION USING LOCAL FEATURE DESCRIPTOR WITH ELM CLASSIFIER Abstract: A hand signal is a nonverbal form of contact between a person and a machine. Because of its future uses in real-life, studies on hand gesture identification are on the rise these days. Computers have been indispensable in everyday activities such as schooling, medicine, entertainment, industry, and digital surveillance. Human-computer interface (HCI) is evolving in contemporary society to improve the smartness of everyday operations. For efficient identification of hand sign images, a novel attribute descriptor is suggested. The production of a vision-based fixed hand sign detection framework for practical uses using a web camera. The following steps are used to create a vision based standardized hand sign recognition system: pre-processing, extraction of features, and labeling. Illumination correction, optimization, filtering, hand area detection, and image resizing are all part of the preprocessing step. In an unregulated setting, the local attribute descriptor extraction method is used to distinguish hand movements. This approach is not only resistant to interference and sign vocabulary, but it is also unobservable to hand sign rotation and translation. To identify hand expressions, an extreme learning machine (ELM) is used as a classification model. This ELM-based approach using a multi-quadric radial base activation mechanism is contrasted to other traditional classification techniques. For hand sign detection, the proposed approach is tested on two common public set of data and one indigenously constructed complex context dataset. The mean precision of the experimental outcomes is assessed. HAND SIGN RECOGNITION USING LOCAL FEATURE DESCRIPTOR WITH ELM CLASSIFIER Diagram INPUT 5 GIVEN AS INPUT S GIVENAS THEHAND SIGN THE HAND SIGN IMAGES IMAGES PRE-PROCESSING PRE-PROCESS]NG EXTRA ON OF EXTRACTiON OF FEATURES FEATURES EXTREME LEARNING MACHINE (ELM) BASED CLASSIFIER USED FOR CLASSIFICARION Figure 1: Block diagram of the proposed hand sign identification system.