Read, Visualize, Evaluate and Analyse

Dr. Priyanka Jain
October 29, 2020

Dr. Priyanka Jain

Automatic Text Visualization (ATV) or Text to Scene (TTS) conversation is relatively a new research area covering the domain of Artificial Intelligence, Natural Language Processing, Cognitive computing, Knowledge engineering, Computer Graphics and Psycho-linguistics. ATV is the process of language understanding of an input language text and its utilization as a 3D scene/ virtual environment. ATV has usage in real time applications like language learning for the persons with specific learning disabilities, knowledge transformation (globalization), text summarization, encrypting information for security and Games & entertainment. There are few works reported in English and other European languages but none for Indian Language.

We are talking about an end-to-end case study for Language Hindi, which is a Morphological rich free word order (MoR-FWO) language with many challenges in its computational processing. Our prior work related to the proposed research is named as Preksha - an Automatic Text Visualizer (ATV) for Hindi language. The Preksha architecture has a user interface to take language input and to produce a 3D virtual environment. The whole processes are supported by a structured resource repository. Preksha has three major computational engines: a Language Engine (LE) for natural language processing of the input text, a Knowledge Engine (KE) to extract and transform the knowledge, and a Scene Engine (SE) for its visualization. We have designed subjective evaluation for the work in absence of standards automatic methodologies on evaluations of ATV systems. Our strategic plan of evaluation carried out to cover Intelligibility, Fidelity and Complexity features of the system. For moderate complex inputs for human to visualize i.e. 1.77/5 the scores are achieved as 3.01/5 i.e. ‘Very Accurate’ and 2.84 /5 ‘Fairly Intelligent’.

The ATV is a relatively new interdisciplinary area. It is challenging because it is not easy to automatically generate a scene which exactly corresponds to the user’s mental image and its expected visualization. It is especially more complex due to the morphological rich and free-word-order nature of Hindi. To the best of the knowledge of the authors, Preksha is the only known work for an Indian Language, viz. Hindi. Belonging to Indian languages, this work on the Hindi Language contributes as a novel interdisciplinary approach to prepare a virtual environment for cognitive support. Visualization provides an environment for learners to actively pursue their knowledge needs by applying their theories in the ‘real world’, therefore, it decreases cognitive load for the person with difficulties in comprehension especially in language reading, e.g. Dyslexia. The research is beneficial in learning a language especially for rehabilitation of a person with some delayed development. Storytelling, news/event visualization are other applications. Gaming and virtual reality environments are expected to be a more promising area of text-to-scene conversion. We poise Interactive Augmented Reality (AR) using natural language to profoundly transform Education to result in high levels of independent thinking, creativity and critical analysis.

Related Posts


Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form