Pradipta Biswas

Mixed Reality (MR) systems integrate digital guidance with physical environments to supportindustrial task execution; however, reliable object detection remains challenging due to occlusion, lighting variation, and limited annotated data. This work addresses data scarcity through diffusion based generative augmentation that expands low-resource industrial datasets with realistic synthetic tool instances under diverse background and illumination conditions, targeting detection failure cases. To validate the effectiveness of the synthetic dataset beyond offline detection metrics, this work conducts a user study in a real MR task execution setting, where detection stability affects the alignment and usability of visual guidance. The study compares three MR assistance modes, Textual Guidance, Bounding Box Guidance, and Holographic Guidance, to assess their impact on user performance, cognitive workload, and usability across varying lighting conditions. Results show that hologram guidance improve visual search time and task performance, with holographic guidancereducing task completion time by 21.4% and workload by 41.9% relative to textual guidance, while increasing usability by 33.4%. These findings demonstrate that diffusion-augmented synthetic data improves object detection robustness and translates into measurable gains in human-centered MR task performance.

Assembling complex components often requires expert guidance. Augmented Reality (AR) offers intuitive visual assistance that can enhance user performance, yet existing AR-based systems largely focus on simple tasks, limiting their application to intricate assembly scenarios. This paper addresses the gap by developing a comprehensive pipeline for AI and AR-based assembly guidance. We analyzed various software architectures, selecting the optimal setup for advanced manufacturing based on latency, accuracy, and human factors. A novel AI-based hologram registration technique was implemented to provide real-time assistance in dynamic environments. Additionally, a multimodal user interface was designed to facilitate seamless interaction with the Mixed Reality (MR) system, allowing users to access instructions efficiently. Results demonstrate that users completed tasks faster with the MR system compared to traditional video-based methods. The proposed system also significantly reduces cognitive load and improves usability, positioning it as an effective tool for modern manufacturing.

In recent years, the internet has emerged as the primary information hub, encompassing various domains such as education, healthcare, government, and e-commerce. With a vast user base relying on websites for accessing essential information, ensuring web content accessibility has become imperative. Meeting accessibility guidelines is crucial to accommodate users with diverse abilities worldwide. This article examines accessibility variances among education board websites hosting higher-level school leaving examinations. The study evaluates the landing pages of 13 educational websites from both developing and developed countries, employing six Web Content Accessibility Guidelines (WCAG) tools endorsed by the World Wide Web Consortium (W3C). The study has meticulously compiled the outcomes of each assessment tool, delineating the diverse accessibility concerns encountered on each website. Furthermore, alternative recommendations have been proposed to assist web developers in rectifying these issues. The findings underscore a higher prevalence of accessibility issues within the education board websites of developing countries compared to their developed counterparts. Specifically, the analysis revealed prominent text contrast failures among selected web pages from developing nations, while non-text contrast failures were more prevalent in those from developed countries. Notably, the education board websites of India and Japan emerged with the highest number of accessibility issues among the sampled websites. As a way forward, a case study was shown on using LLM for making websites accessible.