At Vidhance, Master's thesis projects play a crucial role in driving our exploratory innovation. Each year, we welcome a select group of students to contribute to cutting-edge research in areas that align with our vision for future advancements. Many of our past thesis students have gone on to join our team, continuing their journey with us after completing their projects.

Below are the available thesis topics for Spring 2025. To apply, please click the "apply for this job" button, specifying your preferred topic(s) in your application. Don't forget to include your transcript of records.

1. Quality Testing for Video Understanding LVMs
2. Utilizing Audio Event Input in Automatic Video Editing
3. Exploring Multimodal Interaction for Video Editing: A User-Centered Approach to AI Integration

Quality Testing for Video Understanding LVMs

The objective of this thesis is to develop an objective method for evaluating the quality of video understanding in Large Vision Models (LVMs). The project involves designing and implementing a testing framework that can measure the accuracy and effectiveness of video comprehension across various models. Once developed, this tool will be used to perform comparative analyses of existing video understanding LVMs, providing insights into their strengths and weaknesses. This research will contribute to advancing video analysis technology, particularly in improving the benchmarks for video understanding capabilities.

Utilizing Audio Event Input in Automatic Video Editing

This thesis project focuses on integrating audio event detection into AI-driven video editing workflows. Current AI video editing solutions often prioritize visual analysis, neglecting the contextual significance of audio events like applause, bird songs, and cheering. The project aims to develop a system that generates audio event transcripts from video clips and investigates how these sound cues can enhance the automated editing process. The goal is to ensure that important audio events are included in the final video, improving the overall quality and emotional impact of the edited content.

Exploring Multimodal Interaction for Video Editing: A User-Centered Approach to AI Integration

This thesis project explores innovative interaction methods for AI-powered video editing, moving beyond natural language interfaces. It focuses on multimodal interaction, where users can combine speech, gestures, and visual cues to interact with AI systems. The aim is to design intuitive, user-centered interfaces that require minimal input while still producing high-quality, story-like video edits. By leveraging contextual knowledge from the user’s environment—such as device, location, and preferences—the project seeks to simplify the editing process. Through prototyping and user testing, this research will enhance AI-driven video editing by making it more accessible and engaging for users.