Line Diagram Analysis for Visually Impaired Student - Student Presentation
Join us for an exciting student research presentation!
Come support and learn from students as they share their work, there will be multiple presentations happening throughout the day.
Snacks and lunch will be provided, so bring your curiosity (and your appetite)!
We’d love to see you there!
Agenda:
1:00 pm - 1:25 pm | Line Diagram Analysis for Visually Impaired Student: Ajitesh
1:25 pm - 1:40 pm | Line Diagram Analysis for Visually Impaired Student Q&A: Ajitesh
1:40 pm - 2:00 pm | Snacks & Refreshments: Round Table & Open Discussions
Abstract:
Line diagrams and mathematical function plots are commonly used in scientific and educational textbooks and articles to convey quantitative relationships. However, their visual nature presents significant challenges and barriers for visually impaired learners. Despite the ongoing efforts to improve accessibility in higher education, the non-text content still remains difficult to interpret non-visually. The existing diagram analysis approaches often focus on chart type classification, require manual user intervention for tasks such as label mapping, or fail to provide accessible end-to-end interaction for non visual analysis. Also, many of these systems do not explicitly consider the requirements of visually impaired users, resulting in workflows and interfaces that are not accessible. We propose an accessibility aware framework for automated analysis of line diagrams from raster images. The proposed end-to-end system performs CNN based curve segmentation, diagram type classification, label and tick extraction using optical character recognition, mapping of curve pixels to the diagram domain, and supports interactive querying along with a user interface compliant with the Web Content Accessibility Guidelines. Evaluation is conducted using a large synthetic dataset of 2D chart images representing a variety of mathematical functions and discrete point line charts, which is used for both model training and system-level assessment.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
Loading virtual attendance info...
- 1000 K. L. O. Rd
- Kelowna, British Columbia
- Canada V1Y 4X8
- Building: C
- Room Number: 130
- Click here for Map
- Starts 05 May 2026 07:00 AM UTC
- Ends 07 May 2026 07:00 AM UTC
- No Admission Charge
Speakers
Ajitesh Parihar
Cross-Platform 2D and 3D Way-finding Mobile Application for Navigation in Industrial Buildings and Parking Lots
Line diagrams and mathematical function plots are commonly used in scientific and educational textbooks and articles to convey quantitative relationships. However, their visual nature presents significant challenges and barriers for visually impaired learners. Despite the ongoing efforts to improve accessibility in higher education, the non-text content still remains difficult to interpret non-visually. The existing diagram analysis approaches often focus on chart type classification, require manual user intervention for tasks such as label mapping, or fail to provide accessible end-to-end interaction for non visual analysis. Also, many of these systems do not explicitly consider the requirements of visually impaired users, resulting in workflows and interfaces that are not accessible. We propose an accessibility aware framework for automated analysis of line diagrams from raster images. The proposed end-to-end system performs CNN based curve segmentation, diagram type classification, label and tick extraction using optical character recognition, mapping of curve pixels to the diagram domain, and supports interactive querying along with a user interface compliant with the Web Content Accessibility Guidelines. Evaluation is conducted using a large synthetic dataset of 2D chart images representing a variety of mathematical functions and discrete point line charts, which is used for both model training and system-level assessment.
Biography:
Ajitesh Parihar is a fourth-year Computer Science student and a Research Assistant at Okanagan College. He is a tech enthusiast who enjoys learning about innovations and technologies and is passionate about software engineering, cybersecurity, and algorithms. He enjoys meeting and working with people with diverse backgrounds.
Dolcy Sareen
Activity Diagrams Analysis for Visually Impaired Students
Diagrams are a very important part of education,learning and understanding , allowing the transmission of complex ideas in a visual format.
However, current accessibility technologies introduce substantial limitations for visually impaired students in accessing these materials. Traditional approaches to fix this problem, such as audio descriptions and tactile graphics, either lack detail or are prohibitively expensive to scale. This review will focus on the current status of accessible diagram technologies, such as tactile graphics, 3D printing, AI-driven recognition applications and methods proposed till now . We identify the major technical challenges and propose an integrated approach using multi-modal interaction by combining audio, and AI-enhanced diagram interpretation. The approach offers equal access to STEM visual content as part of the road to inclusive education.
Biography:
Dolcy Sareen is a fourth-year Computer Science student at Okanagan College. Dolcy is also the webmaster of the Okanagan College IEEE Student Branch. She is passionate about learning new technologies, with a strong focus on data. Dolcy enjoys connecting with people and continuously learning from new experiences.