Research Archive

Publications, preprints, and project-linked research outcomes.

This page highlights work across computational imaging, smartphone camera enhancement, atmospheric turbulence restoration, simulation, medical imaging, and vision-driven machine learning. Where external resources are available, direct paper and project links are included on each publication card.

Publications

Selected recent publications

F2IDiff: Real-world Image Super-resolution using Feature-to-Image Diffusion Foundation Model

Generative Imaging 2025

arXiv preprint Recent work at Samsung Research America

F2IDiff: Real-world Image Super-resolution using Feature-to-Image Diffusion Foundation Model

Devendra K. Jangid, Ripon Kumar Saha, Dilshan Godaliyadda, Jing Li, Seok-Jun Lee, and Hamid R. Sheikh

Introduces a feature-conditioned diffusion foundation model for smartphone image super-resolution. The work replaces text-only conditioning with DINOv2 features to improve fidelity, reduce hallucinations, and better support practical camera pipelines.

Paper

DAATSim: Depth-Aware Atmospheric Turbulence Simulation for Fast Image Rendering

Simulation & Rendering 2025

Pacific Graphics / Computer Graphics Forum Published

DAATSim: Depth-Aware Atmospheric Turbulence Simulation for Fast Image Rendering

Ripon Kumar Saha, Yufan Zhang, Jinwei Ye, and Suren Jayasuriya

Presents a fast, physically grounded atmospheric turbulence simulator that combines depth estimation with wavefront-inspired rendering to generate spatially varying blur, geometric distortion, and temporally consistent degradation.

Paper Project Site

Full publication list

Multimodal Vision

MetaVIn: Meteorological and Visual Integration for Atmospheric Image Degradation Estimation

Saha, Ripon Kumar, Mccloskey S, Jayasuriya S

WACV 2025 Conference submission

Combines meteorological cues and visual features to estimate atmospheric image degradation more robustly than image-only baselines.

Restoration

Turb-Seg-Res: A Segment-then-Restore Pipeline for Dynamic Videos with Atmospheric Turbulence

Ripon Kumar Saha, Qin D, e J, Li N, and Jayasuriya S

CVPR 2024 Published

Introduces a dynamic-scene restoration pipeline that separates motion understanding from restoration to improve turbulence mitigation in complex videos.

Segmentation

Unsupervised Region-Growing Network for Object Segmentation in Atmospheric Turbulence

Qin D, Saha, Ripon Kumar, Jayasuriya S, Ye J, and Li N

ECCV 2024 Published

Presents an unsupervised segmentation framework tailored for distorted turbulent imagery where stable object boundaries are difficult to recover.

Physics-based ML

Turbulence Strength C2n Estimation from Video using Physics-based Deep Learning

Ripon Kumar Saha, Esen S, Jihoo K, Joseph S, and Suren J

Optics Express 2022 Published

Develops a physics-based deep learning framework for estimating atmospheric turbulence strength directly from video in real imaging conditions.

Medical Imaging

Automated Quantification of Meibomian Gland Dropout in Infrared Meibography using Deep Learning

Ripon Kumar Saha, Chowdhury AM, Na KS, Hwang GD, Hwang H, and Chung E

Ocular Surface 2022 Published

Uses deep learning for automated gland analysis and clinical grading from infrared meibography, enabling robust ocular surface assessment.

Vision & Perception

Using a CNN Model to Assess Visual Artwork's Creativity

Zhehan Zhang, Meihua Qian, Li Luo, Ripon Kumar Saha, Qianyi Gao, and Xinxin Song

APA / arXiv 2024 Published / preprint

Explores CNN-based modeling of perceived creativity in visual artwork, connecting computer vision features with psychological evaluation.

Biomedical AI

Electrocorticography-Based Motor Imagery Movements Classification using LSTM based on Deep Learning Approach

Md Rashid, Md Islam, N. Sulaiman, B. S. Bari, Ripon Kumar Saha, and M. J. Hasan

SN Applied Sciences 2020 Published

Applies LSTM-based modeling to electrocorticography signals for motor imagery movement classification in a biomedical setting.