Ripon Kumar Saha, Arizona State University, Computer Vision and Computational Photography

About Me

I am currently a Ph.D. student at Arizona State University, specializing in Computer Vision, Machine Learning, and Computational Photography within the Imaging Lyceum Lab, under the guidance of Dr. Suren Jayasuriya. My research is focused on developing advanced deep learning models for dynamic scene restoration, particularly addressing the challenges posed by atmospheric turbulence in Ultra-Zoom and astrophotography applications. This work integrates physics-based methods with SOTA Deep learning techniques, aiming to significantly enhance image clarity and stability in challenging imaging environments.

Prior to my Ph.D., I completed an M.S. in Biomedical Science & Engineering at the Gwangju Institute of Science and Technology (GIST) in South Korea, where I worked as a research assistant in the NeuroPhotonics Lab under Professor Euiheon Chung. My thesis involved developing a deep learning model for the automated assessment of infrared images of the tear film, including segmentation of the Meibomian gland and specular reflection removal. This work resulted in the release of a 1,000-image annotated dataset and demonstrated advanced techniques using ResNet50 and GANs.

I earned my Bachelor’s degree in Computer Science & Engineering in Bangladesh, where I developed strong skills in C++, Python, web design, Java, and iOS app development with Swift. I actively participated in national and regional programming contests and hackathons, honing my problem-solving abilities and gaining recognition in various competitive events.

Contact Details

Ripon Kumar Saha
Arizona State University
Tempe, Arizona, USA
602 -802- 9345
ripon.ece@gmail.com

Education

Arizona State University (ASU)

PhD student with RA in Imaging Lyceum Lab August 2020 - Containue

My current work is based on Turbulence Image Restoration from dynamic scenes using self-supervised deep learning methods. Overall I'm working on Computational Photography and deep learning area. Besides, I'm working for Alphacore related to the Turbulence project and LightSense Technology related to spectroscopy in collaboration with the Imaging Lyceum Lab supervised by professor Suren Jayasuriya.

Relevant Coursework:

  • EEE 515 Machine Vision and Pattern Rec
  • EEE 598 Physics-based Computer Vision
  • CSE 551 Foundations of Algorithms
  • EEE 350 Random Signal Analysis

Gwangju Institute of Science and Technology (GIST)

Master student with RA in Neurophotonics Lab September 2018 - August 2020

My current lab mainly focuses on solving current outstanding problems using state-of-the-art neurophotonics technology. My work is based on computational image analysis based on biomedical images, and testing different ML algorithms, model optimization & developing new models.

Relevant Coursework:

  • Computer Vision
  • Deep Learning
  • Advanced Deep Learning
  • Algorithm
  • Fast.ai Mooc 2019
  • CNN (Stanford Online)
  • The Ancient Secrets of Computer Vision (UW online - ongoing)
  • Deep Learning Specialization (By Coursera - ongoing)

Jashore University of Science and Technology

B.S.C. Degree in Computer Science & Engineering February 2012 - December 2016

For thesis, I designed real-time Earliest deadline first algorithm with Relatively Greater Power Efficiency and Low Deadline Miss Ratio. I have also Completed various projects including Bengali language speaker in C#, Assignment maker in C++, and OpenGL GUI game of bubble shooter.

Industry Experience

Kitware Inc.

Summer Internship May 2024 - Aug 2024

Applied deep learning and computer vision methods for object detection, event/activity recognition, and video/image understanding. Calculated uncertainty in object recognition and detection in long-range video footage. Utilized data from ground, handheld, aerial, or satellite cameras, advancing real-time segmentation, enhancing multiple degraded videos, optimizing performance, and emphasizing object feature preservation.

Lightsense Technology Inc.

Summer Internship June 2022 - August 2022

Developed an AI model for Covid-19 classification using spectral data and machine learning techniques. Pioneered spectral unmixing solutions for bacteria samples, analyzing viruses in saliva and buffer solutions using the PARAFAC algorithm and various preprocessing techniques. Enhanced component identification vital for drug detection, pathogen identification, food safety, and medical diagnostics.

Alphacore Inc.

Doctoral Student Collaboration March 2021 - August 2023

Managed onsite field experiments setup with several telescopes, drones, cameras, weather stations, and scintillometers. Built a deep learning model for atmospheric turbulence estimation across varying focus distances, light intensity, platform motion, and camera shake. Analyzed and processed extensive multidimensional data from various sensors, contributing to research advancements in atmospheric turbulence estimation.


Publications

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

Unsupervised Region-Growing Network for Object Segmentation in Atmospheric Turbulence

Qin D, Ripon Kumar Saha, Jayasuriya S, Ye J, and Li N ECCV 2024

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

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

Using a CNN Model to Assess Visual Artwork's Creativity

Zhehan Zhang, Meihua Qian, Li Luo, Ripon Kumar Saha, Qianyi Gao, and Xinxin Song arXiv preprint arXiv:2408.01481, American Psychological Association (APA), 2024

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

Rashid M, Islam M, Sulaiman N, Bari BS, Ripon Kumar Saha, Hasan MJ SN Applied Science 2020


Skills

My research interests include ML-based computer vision, computational photography, model design & optimization, stereo vision, and bioinformatics. Below is a list of my relevant skills:

  • Programming Languages: Python, MATLAB, C, C++, Java, SQL, Bash, HTML5, Swift, PHP, JavaScript, CSS
  • Frameworks & Libraries: PyTorch, TensorFlow, Keras, Fast.AI, Pandas, NumPy, Scikit, NLTK, OpenCV, Flask, J2EE
  • Machine Learning & AI: CNN, FCN, RNN, LSTM, Diffusion Models, GAN, Transformers, Large Language Models
  • Data Visualization and Analysis: Tableau, Microsoft PowerBI, Seaborn, Origin-Pro, GraphPad
  • High-Performance Computing: Batch Scripting, GPU Clusters, Python Multi-Processing, Dask, Cython
  • Version Control and DevOps: Git, Docker, MySQL
  • Computational Optics & Microscopy: Experience with developing optical microscopy systems and computational optics.

Achievement

My Achievement includes:

  • 1st place in BuildwithAI Hackathon [4,000+ participants, 300+submission, 70+ countries].
  • 1st Runners-up in the National Math Olympiad Bangladesh. [5400 participants].
  • Champion within 6 districts in the Regional math olympiad.
  • Received South Korean Government Scholarship for MS.
  • Rated “Specialist” in Contest Programming Platform Codeforces.

PROJECTS

Mejor projects I have worked on:

  • Multimodal Deep learning architecture with GAN impainting and encoder-decoder based network for segmentation and qualitative analysis of Meibomian Gland [achieved better performance than Ophthalmologist]
  • Single snapshot blood speed detection with the physics-based property.
  • Image analysis to detect blood glucose from a contact lens.
  • Developing “Abbe Diffraction microscopy,” “Confocal microscopy.”
  • Dark image denoiser with memory optimization.
  • Construct 3D structure from multiple viewport images based on own created dataset with camera calibration.
  • Construct All in-focus images from multiple defocused images.
  • Time series Media bias & sentiment on COVID-19 [twitter scraping data with BERT].

Tear Film Diagnosis with Deep Learning

The Ocular Surface (In Review)

Developed a multimodal deep learning model for diagnosis of Meibomian Gland (paper)

Tear Film Diagnosis with Deep Learning

The Ocular Surface (In Review)

Developed multimodal architecture for automated assessment of tear film infrared images to detect/segment out the eye gland area, provide ophthalmologist quality assessment score (Meiboscore) and remove specular refection.

Dataset of 1000 images released.

[Model: Encoder-Decoder Structure, Resnet50, GAN architecture with image inpainting].

Image Reconstruction in Turbulence

Current Lab Project

Designing a physics-based deep learning model for dynamic scene restoration affected by atmospheric turbulence taken with Ultra-Zoom or astrophotography camera.

Image Reconstruction in Turbulence

Current Lab Project

Images taken by long distance object object is effected by turbulenec and come out distorted and blurry. We are working on image restoration from dynamic moving scene of different depth using self supervised deep learning approach.

This involve moving object separation from turbulence motion by trained optical flow network and self supervised way of image restoration.

Path Averaged Structure Index Estimation

Lab Project

Design ML model for atmospheric turbulence estimation with focus, light and motion correction.

Path Averaged Structure Index Estimation

Lab Project

Onsite Experiment: Setup onsite team experiment with several telescopes, weather stations, and scintillometers.

Data Analysis: Analyze data taken with telescope, drones, cameras, weather stations, and scintillometers.

ML Model: Design ML model for atmospheric turbulence estimation with focus, light and motion correction.

Covid-19 Detection from Spectral Signature

Lab Project - with Collaboration

Develop a AI model for classifying different varieties of COVID virus by analyzing spectral data.

Covid-19 Detection from Spectral Signature

Lab Project

Spectral Analysis: Analyze absorption and emission spectroscopy data of viruses from saliva and buer solution

Covid-19 Classication: Simulate dataset from limited spectra; AI for Covid-19 classication from spectral signatures.

Funded by: Lightsense Technology

Detect Blood Glucose from Contact Lens

Ms Project

Developed a architecture to analyze color of images of contact lenses and predict blood glucose level.

Detect Blood Glucose from Contact Lens

Ms Project

Developed an architecture to analyze images of custom contact lenses and predict blood glucose level with 85% accuracy

Measured with spectroscopy system and got better result from image analysis based methods.

Building Optical Microscopy/Telescope

MS Project

I assisted developing different optical microscopy system in the Lab.

Building Optical Microscopy/Telescope

MS Project

I built abby's diffraction limited microscopy consisting of two camera and different lens elements with supervision from Muhammad Mohsin Qureshi and Professor Euiheon chung.

Assisted lab members building different optical system including Confocal microscopy(with integration of scanimage), Light sheet microscopy and two Photon Microscopy.

Motor Imagery Movements Classification with LSTM

Paper

Classifying imaginary hand movement activities from ECG signal.

Motor Imagery Movements Classification with LSTM

Paper

We have proposed a LSTM based deep learning approach where no need to employ any feature extraction and reduction framework. In this approach, raw ECG data have been employed to the LSTM model.

Two classes motor imagery ECG has been classified using LSTM based deep learning approach.

COVID 19 Political Analysis from 387K Scrapped Twitter

Hackathon

Classifying imaginary hand movement activities from ECG signal.

COVID 19 Political Analysis from 387K Scrapped Twitter

Hackathon

Utilize NLP to assess the tweets of influential personalities and how it influences public sentiment
Understand public perception and the influence of misinformation on public sentiment to

  • Help Policymakers with data-driven decision making to check spread of misinformation.
  • Inspire Media and influential personalities to be cognizant and share the right news.
  • Wake-up General Public to be more conscious about their news consumption.

Find Generalized Neural Network for Tabular Data

MS Coursework

Classifying imaginary hand movement activities from ECG signal.

Find Generalized Neural Network for Tabular Data

MS Coursework

This project was focused on measuring different neural network model available today and find the best suited model for tabular format data.

12 data sheet was collected from kaggle for Evaluation

Compared SVM, Attention Net, ResNet, CNN, Attention+Resnet with different learning rate, optimizer, loss function, decay, dropout, cross validation to find the best neural network basnd on One hot encoding was used.

Recommandation

  • I'll update Quote one I receive recommandation from someone.

    Name of the person
  • I'll update Quote one I receive recommandation from someone else.

    Name of the person 2

Get In Touch.

If you want to connect me, you can seme me an email here. If you want to connect me, you can seme me an email here. I'm open to new research/project ideas. I like to connect people interested in computer vision, deep learning, and computational photography. If you need any quick help/advice/consultation regarding your custom computer vision project, I'll be happy to provide guidance.