Swarna Chakraborty
Texas State University
I am a fourth-year Ph.D. candidate at Texas State University, advised by Dr. Mylène C.Q. Farias. My research focuses on perceptual quality assessment of immersive and volumetric media, using vision-language models to understand how humans perceive 3D content.
I am graduating in Spring 2027 and actively exploring postdoctoral positions in closely related fields. I am also open to research-track roles in industry.
Before starting my Ph.D., I completed my bachelor's degree in Computer Science & Engineering at Chittagong University of Engineering and Technology (CUET), Bangladesh and spent two years as a software engineer at Enosis Solutions.
Research
My research sits at the intersection of human perception, computer vision and multimodal learning. I build models that predict how people experience the quality of immersive and volumetric media like 3D point clouds, dynamic scenes, and other non-standard modalities where traditional image metrics break down with a focus on explainability and interpretability. A central thread of my work is adapting large vision-language models to reason about geometry and distortion, so that quality assessment can be both accurate and explainable.
Research interests
Publications
DAL-PCQA: Enabling Distortion-Level and Language-Driven Reasoning for Point Cloud Quality Assessment
2026 18th International Conference on Quality of Multimedia Experience (QoMEX)
Proceedings of the 33rd ACM International Conference on Multimedia, pp. 7113–7122
A No-Reference Point Cloud Quality Assessment Using Graph Attention Networks and Keypoint Resampling
13th European Workshop on Visual Information Processing (EUVIP '25)
Technical Skills
Languages
- Python
- C / C++ / C#
- Java
- SQL
ML / DL Tools
- PyTorch
- TensorFlow
- OpenCV · Open3D
- Scikit-learn
- NumPy · Pandas
- Transformers / HuggingFace
Frameworks
- .NET Core
- Django
- AngularJS
- jQuery
Experience
Doctoral Instructional Assistant
Texas State University · San Marcos, TX
Conduct laboratory sessions for undergraduate CS courses. Explain programming concepts, provide hands-on support during exercises, and grade assignments. Mentor students in foundational computational thinking.
Software Engineer
Enosis Solutions · Bangladesh
Developed large-scale cloud-based applications using MVC and onion architecture. Conducted requirement analysis, feature implementation, and bug resolution with .NET Core, C#, jQuery, AngularJS, and SQL Server.
Education
PhD in Computer Science
Texas State University · GPA 4.0/4.0
B.Sc. in Computer Science & Engineering
Chittagong University of Engineering and Technology · GPA 3.46/4.0
Selected Projects
01 —
GitHub ↗Jute Pest Classification via Deep Learning
Developed and benchmarked CNNs, Vision Transformers, and ResNet hybrids for classifying jute pest images. Achieved 98% accuracy through systematic experimentation and hyperparameter tuning.
02 —
GitHub ↗Argument Mining via NLP Transformers
Designed transformer-based models (RoBERTa, Swin Transformer) for automated argument mining. Applied SMOTE for class imbalance and PyTorch fine-tuning to maximize precision, recall, and F1 across argument components.
03 —
GitHub ↗Evaluating User Experience in VR-Based Charades: A Comparison with Traditional Digital Platforms
Evaluated user experience in a VR-based charades game, comparing interaction quality, engagement, and usability against traditional digital platforms like Microsoft Teams through experimental analysis.
04 —
GitHub ↗Perceptual Quality Assessment of AI Generated Images
Benchmarked CNN, ResNet-50, and Vision Transformer architectures against human MOS ratings on multiple AIGIQA datasets. Then designed ViT-MSQA, a multi-scale, quality-aware attention model that outperformed several published IQA methods.
Service
Reviewer for the following venues:
- Reviewer IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
- Reviewer ACM International Conference on Multimedia (ACM Multimedia)
CV
Curriculum Vitae
Full academic CV — publications, teaching, and research experience (PDF)
Contact
Let's talk research, collaborations, or opportunities.