IMAGIBOT – An Image Recognition Chatbot for Sri Lankan Ancient Places

Abstract

Ancient artifacts of Sri Lanka are used as major sources of getting historical information about Sri Lanka. Acquiring the necessary information about those artifacts become a huge challenge for the visitant of artifacts. Text based search engines are typically used to retrieve information about ancient artifacts of Sri Lanka. These systems require the user to formulate a text query that provides information such as the place, where the object is placed or where does that can be found, what is the century object belonged. Oppositely visual search systems can be used. They provide information to users of the system such as scholars, tourists, local explores in a most intuitive and immediate way by using an image as a query. This research involves in developing a chat application along with Convolutional Neural Networks for image recognition of ancient artifacts. The image recognition model will be a part of the chat bot that has ability to retrieve more information about recognized images. Convolutional Neural Networks, Recurrent Neural Network, TensorFlow, Keras have been used as core technologies in this research project. As the research involves in developing a chat application with image recognition capabilities, the application will improve the knowledge sharing of Sri Lankan ancient legacy to the word in an effective way.