Design and development of a tangible machine learning-based object recognition and learning system for Indian kids suffering from visual agnosia

Children with visual agnosia have problems recognising and remembering basic objects. Although they are not visually impaired, they may be disregarded in educational environments. They suffer from the issue of limited identification capacity and are unable to retain what they have seen. A physical device using machine learning has been developed to assist individuals with difficulties in recalling and recognising basic objects like fruits and vegetables. The gadget can recognise objects and provide detailed descriptions within a 100-word limit. The designed equipment aids children in retaining knowledge of several fundamental items and enhances their inquisitiveness towards them. Consequently, they are able to gradually develop their cognitive abilities, including sensorimotor and object permanence capabilities. The object detection is performed via the convolutional neural network (CNN) method implemented in TensorFlow. The equipment effectively conducted tests on more than 20 children, examining a minimum of 30 distinct items.