The Power of Visual Storytelling: A Deep Learning Framework for Educational Influence

Abstract

This research focuses on enhancing students' memory retention and visualization skills in education through visual storytelling, a powerful pedagogical approach known for improving learning outcomes across various domains. Our visual storytelling method involves generating images from text using Generative Adversarial Networks (GANs) along with the Graph Embedding technique, contributing to comprehension, memory, sequencing, and engagement. A primary survey revealed that a significant portion of teachers (70.3%) in India and Cambodia incorporate visual storytelling in half of their total class time. However, the process is perceived as time-consuming, prompting educators' expectations for a unified model capable of generating complete reading comprehension along with images. In our post-experiment test, involving 8th-grade students, the study compared the Experimental Group, experiencing visual storytelling, with the Control Group using traditional methods. Assessments through Fill-in-the-Blanks, Multiple Choice Questions (MCQs), and story continuation demonstrated the Experimental Group's improved recall, application, and critical thinking skills. Remarkably, 92% of the Experimental Group actively engaged in continuing the story, signifying a substantial increase compared to the Control Group. Post-research surveys show a strong willingness of educators (96.3%) to integrate visual storytelling into classrooms, with 63% expressing a desire to incorporate our research findings into teaching methods, particularly in reading comprehension with image sequences. These findings underscore the positive impact of visual storytelling, advocating for its widespread adoption and further exploration in educational practices.