Page 234 - AWSAR 2.0

210 || AWSAR Awarded Popular Science Stories - 2019
belief to me as if abstract visualization can support the information transformation to break the language barrier in a multilingual country like India.
The belief was fortified when my stories to my toddler got shaped as the sketches instead of vocal form. In her primary school, the visual representation of tales was easy for her to capture rather than reading these texts in the books. Most of the time we had fun with our conversation in pictorial forms. I not only observed but studied
also that visual data has higher
bandwidth which reduces
cognitive load in understanding.
Persons create mental images
while and after they read. The
theoretical cognitive process
of visualization for science
education is covered by some
of the researchers.
Based on these
approaches, I initiated the
research Preksha - a Hindi text
visualizer under the guidance
of Dr. Hemant Darbari, Director
General, C-DAC and Dr. B.
V. Pawar, Professor, KBC-NMU, Jalgaon. The automatic text visualization (ATV) is a process of deliberately constructing a scene from natural language text to meet specified communicative goals. As an aid for language learning, our research is focused to process Hindi language, extendible to other similar
Indian languages. As the earlier reported works for ATV pertain to English or other European languages, therefore, with the desire to scientifically verify the problem especially with the Hindi language. Hindi is found a useful language to investigate since it has certain properties like morphological richness and free-word-order nature which were not seen in previously studied languages for visualization purposes.
ATV is a newer interdisciplinary research area, based on the work in natural language processing (NLP), knowledge engineering (KE) and computer graphics (CG). One of the main issues in text comprehension is that not every text can be visualized. The narrative text, which contains stories, event descriptions, places depictions, news commentaries, action behaviours, architectures and designs is easy to visualize in some cases, whereas non- fictional, purely scientific and philosophical
texts are hard to show in a picture form. The visualizability of text also depends on the imagination and visualization power of the perceiver as an individual. In this research, we coined the concept ‘degree of visualizability (DoV)’ as a parameter to measure the visualizability in text. The DoV can be calculated based on our concept of Object Visualization Feature sets (OVFs) which we have extended for visualization purposes from grammatical semantic feature sets. There
are some vital issues in automatic text-to- scene conversion systems like the appropriate comprehension of the input language, context disambiguation, lexicon and ontology preparation, common sense knowledge transformation and spatial relation resolution.
Preksha architecture comprises a
Preksha architecture comprises a consolidated framework with resource repositories, processing engines, computational tools, and with a mechanism of data and control flow. This framework supports tools for resource management for linguistic and scene repositories.