COLLECTION OF DATASETS AND APPLICATION OF NEURAL NETWORK MODELS FOR SIGN LANGUAGE CLASSIFICATION IN PATTERN RECOGNITION TASKS

Abstract

Currently, more and more research are aimed at solving problems using computer vision libraries and artificial intelligence tools. The most common are solutions and approaches using machine and deep learning models of artificial neural networks for recognizing gestures of the Kazakh sign alphabet based on supervised learning and deep learning methods for processing sequential data. The object of the study is the Kazakh sign alphabet for building communication between people with disabilities. The subject of the study is machine learning methods and models of artificial neural networks and deep learning for classifying and recognizing gestures. Research methods are Data Science, Machine Learning, Deep Learning, Neural networks and Computer Vision.

Pattern recognition is an image on which an object is located. Since the object is abstract (the object can be any shape depicted in the picture). We decided to explore one of the current areas - gesture recognition. To recognize the Kazakh sign language, first, you need to learn the Kazakh sign alphabet. To train a neural network to recognize Kazakh sign language, it is necessary to collect data (datasets) in the format of images designated by hand gestures. Gesture recognition is a classification task, which is one of the areas of pattern recognition. The fundamental basis of recognition is the theory of pattern recognition.