Mylene Queiroz De Farias
Texas State University, TX·Computer Science
When several people meet inside a shared virtual world and see three-dimensional recordings of each other, how good does that imagery actually look to them? Measuring visual quality automatically is a solved problem for ordinary flat video, but those tools break down on the three-dimensional content that drives modern virtual reality (VR).
- Fields
Computer Science
Engineering
- Topics
virtual reality
volumetric video
point cloud quality assessment
- Grants
- 1 known grant, through Jun 2028
What they work on
- Design a feature-extraction architecture for dynamic point clouds that jointly captures geometry, texture and motion, using visual transformers and multimodal fusion with temporal awareness and adaptive projection strategies.
- Build a first-of-its-kind multi-user volumetric video quality dataset of real scenes with multiple people and objects, annotated with mean opinion scores, NASA-TLX cognitive workload and Simulator Sickness Questionnaire ratings.
- Collect objective viewer data in high-fidelity head-mounted displays, including eye-tracking and field-of-view traces, to capture how user movement and viewing direction influence perceived quality in shared virtual spaces.
- Train and evaluate a multimodal point cloud quality assessment model that integrates a visual saliency predictor and spatial-temporal fusion, testing predictions against human ratings with correlation, error and ablation analyses.
More about the research
Without a reliable quality measure, developers cannot tell whether compression, transmission or capture choices are degrading the experience. This project builds new automated ways to score the quality of volumetric video, which represents people, objects and whole scenes in three dimensions, with attention to how it is perceived when multiple users share and move through the same virtual space.
Why it matters
Virtual reality is being used for remote collaboration, healthcare, education and entertainment, and its usefulness depends on whether users perceive the content as high quality. Reliable automated quality metrics for three-dimensional volumetric video could let developers tune compression and streaming without degrading the experience, and the planned public dataset of human ratings, eye-tracking, workload and motion-sickness measures gives other researchers a shared benchmark for immersive media.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Tian Guo
Worcester Polytechnic Institute, MA·Computer Science·Early-career
Design QCloud, a quantum cloud framework with job placement, job scheduling and network scheduling algorithms built on models of quantum circuits, gates and remote gates executed across networked machines.
Neil Charness
Weill Medical College of Cornell University, NY·Psychology
Identify older adults' preferences, usability problems and training needs for the immersive VR program CAST (Cognitive Activity Social Technology), using focus groups, usability testing and heuristic analysis across three CREATE sites.
Yuhao Zhu
University of Rochester, NY·Computer Science·Early-career
Develop the foundations for in-sensor machine vision that combines computational metasurfaces (flat nanostructures that process light) with analog encoders inside CMOS image sensors, so images are partly processed before they are digitized.
The school
Texas State University: 8 labs on PI Pages, $165M a year on research, 71 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.