M. Shifat-E-Rabbi
Past - Research Assistant - Imaging and Data Science Lab. DSP research lab.
Past - BSc in Electrical and Electronic Engineering (EEE) from Bangladesh University of Engineering and Technology (BUET), Bangladesh.
Past - Rangpur Cadet College, Rangpur Zilla School, Bangladesh.
Overarching Direction
Mathematical modeling, machine learning, data science, computational biology, applied mathematics, computer vision, pattern recognition.
"Let the future tell the truth, and evaluate each one according to his work and accomplishments. The present is theirs; the future, for which I have really worked, is mine." - Nikola Tesla (1856 – 1943)
Recently published: It is first useful to view an image class as a single "template" image that has been altered by one or more confounding factors to produce the other images in the class. If these alterations can be appropriately modeled as a set of smooth, nonlinear transformations, then different image classes become easily separable in the R-CDT space via the properties we derive in our paper. These properties also allow for the approximation of image classes as convex subspaces in the R-CDT space, providing a more appropriate data model for the nearest subspace method.
The number of classification problems that can be viewed this way is actually quite large. Heuristically speaking, any problem for which one image in a class can be constructed from another by a smooth rearrangement of pixel intensities is a good candidate. Obvious examples are translation and scaling. A less obvious example is distortions due to the influence of a transparent medium.
Our method performs equivalent to or better than the deep learning models (as well as other models), in both low and high data regimes, with up to 10,000 times savings in the computational cost. Besides, the method is also robust in challenging practical scenarios, e.g., the out-of-distribution setup, meaning our model generalizes to previously unobserved data.
Journal articles
Shifat-E-Rabbi M, Pathan NS, Li S, Zhuang Y, Rubaiyat AHM, Rohde GK. Linear optimal transport subspaces for point set classification. submitted to JMIV.
Naamani KE, Rizvi T, Shifat-E-Rabbi M, Kundu S, Becerril-Gaitan A, Chen CJ, Mayer S, Connolly E, Rohde GK. Optimal mass transportat for fully automated hematoma expansion prediction from non-contrast computed tomography in spontaneous intracerebral hemorrhage patients. submitted to Nature Neuroscience.
Gong L, Li S, Pathan NS, Shifat-E-Rabbi M, Rohde GK, Rubaiyat AHM, Thareja S. The radon signed cumulative distribution transform and its applications in classification of signed images. preprint.
Shifat-E-Rabbi M, Ironside N, Ozolek JA, Singh R, Pantanowitz L, Rohde GK. Transport-based morphometry of nuclear structures of digital pathology images in cancer. submitted to Bioinformatics.
Zhang C, Herbig M, Zhou Y, Nishikawa M, Shifat-E-Rabbi M, Kanno H, Yang R, Ibayashi Y, Xiao TH, Rohde GK, Yatomi Y, Goda K. Real-time intelligent classification of COVID-19 and thrombosis via massive image-based analysis of platelet aggregates. Cytometry Part A. 2023 Jun;103(6):492-9. (arxiv paper)
Rubaiyat AHM, Li S, Yin X, Shifat-E-Rabbi M, Zhuang Y, Rohde GK. End-to-End Signal Classification in Signed Cumulative Distribution Transform Space. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024 Mar 1. (paper, arxiv paper)
Zhuang Y, Li S, Shifat-E-Rabbi M, Yin X, Rubaiyat AH, Rohde GK. Local Sliced-Wasserstein Feature Sets for Illumination-invariant Face Recognition. arXiv preprint arXiv:2202.10642. 2022 Feb 22. (arxiv paper) submitted to Pattern Recognition.
Shifat-E-Rabbi M, Zhuang Y, Li S, Rubaiyat AH, Yin X, Rohde GK. Invariance encoding in sliced-Wasserstein space for image classification with limited training data. Pattern Recognition. 2022 Dec 15. (paper, arxiv paper, software)
Zhou Y, Nishikawa M, Kanno H, Xiao T, Suzuki T, Ibayashi Y, Harmon J, Takizawa S, Hiramatsu K, Nitta N, Kameyama R, Peterson W, Takiguchi J, Shifat-E-Rabbi M, Zhuang Y, Yin X, Rubaiyat AHM, Deng Y, Zhang H, Rohde GK, Iwasaki W, Yatomi Y, Goda K. Massive image-based single-cell profiling reveals high levels of circulating platelet aggregates in patients with COVID-19. Nature Communications. 2021 Dec 9;12(1):1-2. (paper, arxiv paper)
Shifat‐E‐Rabbi M, Yin X, Rubaiyat AHM, Li S, Kolouri S, Aldroubi A, Nichols JM, Rohde GK. Radon cumulative distribution transform subspace modeling for image classification. Journal of Mathematical Imaging and Vision. 2021 Aug 5:1-9. (paper, arxiv paper, software)
Zhou Y, Nishikawa M, Kanno H, Xiao T, Suzuki T, Ibayashi Y, Harmon J, Takizawa S, Hiramatsu K, Nitta N, Kameyama R, Peterson W, Takiguchi J, Shifat-E-Rabbi M, Zhuang Y, Yin X, Rubaiyat AHM, Deng Y, Zhang H, Rohde GK, Iwasaki W, Yatomi Y, Goda K. The landscape of circulating platelet aggregates in COVID-19. medRxiv. 2021 Jan 1. (arxiv paper)
Kundu S, Ashinsky BG, Bouhrara M, Dam EB, Demehri S, Shifat-E-Rabbi M, Spencer RG, Urish KL, Rohde GK. Enabling early detection of osteoarthritis from presymptomatic cartilage texture maps via transport-based learning. Proceedings of the National Academy of Sciences. 2020 Sep 21. (paper, software)
Shifat‐E‐Rabbi M, Yin X, Fitzgerald CE, Rohde GK. Cell Image Classification: A Comparative Overview. Cytometry Part A. 2020 Feb 10. (paper, arxiv paper, software)
Islam MS, Shifat-E-Rabbi M, Dobaie AM, Hasan MK. PREHEAT: Precision heart rate monitoring from intense motion artifact corrupted PPG signals using constrained RLS and wavelets. Biomedical Signal Processing and Control. 2017 Sep 1;38:212-23. (paper)
Shifat-E-Rabbi M, Hasan MK. Speckle tracking and speckle content based composite strain imaging for solid and fluid filled lesions. Ultrasonics. 2017 Feb 1;74:124-39. (paper)
Hasan MK, Shifat-E-Rabbi M, Lee SY. Blind deconvolution of ultrasound images using l1-norm-constrained block-based damped variable step-size multichannel LMS algorithm. IEEE transactions on ultrasonics, ferroelectrics, and frequency control. 2016 Jun 7;63(8):1116-30. (paper)
Conferences
Rubaiyat AHM, Shifat‐E‐Rabbi M, Zhuang Y, Li S, Rohde GK. Nearest Subspace Search in The Signed Cumulative Distribution Transform Space for 1D Signal Classification. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE. 2022, Singapore. (arxiv paper, software)
Shifat-E-Rabbi M, Rohde GK. Scientific tutorial at CYTO conference, 2019 in Vancouver, Canada -- Cell Image Classification: An Overview of Methods with Software Examples. (abstract, slide)
[Abstract] Shifat-E-Rabbi M, Rubaiyat AM, Zhuang Y, Rohde GK. A sliced-Wasserstein distance-based approach for out-of-class-distribution detection.
[Abstract] Zhang C, Herbig M, Zhou Y, Nishikawa M, Shifat-E-Rabbi M, Kanno H, Yang R, Ibayashi Y, Xiao TH, Rohde GK, Sato M, Kodera S, Daimon M, Yatomi Y, Goda K. Real-time intelligent classification of COVID-19 and thrombosis with label-free bright-field images of platelet aggregates. JSAP 2023.
Ongoing works
UVa and Vanderbilt crew. Signed R-CDT. (ongoing work)
UVa and DSA crew. Learning from spatial information in sports using Transport Based Morphometry. (ongoing work)
UVa and pathologists. Cancer morphometry. (ongoing work)
UVa, University of Tokyo, and cytometrists. High dimensional distribution classification. (ongoing work)
Face detection via Wass distance
"If you can't explain it simply, you don't understand it well enough." - Albert Einstein (1879 - 1955)
Thesis/Dissertation
Ph.D. Dissertation, BME, UVa, May 2023. M Shifat-E-Rabbi. Transport Generative Models in Pattern Analysis and Recognition. (PhD dissertation)
B.Sc. Thesis, EEE, BUET, Sep 2015. M Shifat-E-Rabbi. Composite strain imaging for solid and fluid-filled lesions. (undergraduate thesis)
Miscellaneous
Software developments: one of the contributors in PyTransKit , TBM package.
Kundu S, Ashinsky BG, Bouhrara M, Dam EB, Demehri S, Shifat-E-Rabbi M, Spencer RG, Urish KL, Rohde GK. Reply to Roemer and Guermazi: Early biochemical changes on MRI can predict risk of symptomatic progression. Proceedings of the National Academy of Sciences of the United States of America. 2021 Mar 16;118(11):e2024679118. (reply letter)
A Aldroubi, S Li, GK Rohde. Partitioning signal classes using transport transforms for data analysis and machine learning. 2020. Co-featured in Acknowledgement. (arxiv paper)
M Shifat‐E‐Rabbi, X Yin, CE Fitzgerald, GK Rohde. Featured in a short highlight on the cover article for Cytometry Part A. Apr 2020 issue. (cover article)
MK Hasan, M Shifat-E-Rabbi, SY Lee. Cover article for IEEE transactions on ultrasonics, ferroelectrics, and frequency control. Aug 2016 issue. (cover article)
Selected works:
(1) Cell image classification is a computational approach for determining the class or category of individual cells from image data information. Cell image classification methods can be used in numerous applications in cell biology and medicine, including understanding the impacts of genes or drugs, learning the roles of subcellular proteins or staining patterns, as well as diagnosis and prognosis of certain diseases. Our article describes three main approaches currently in use for classifying images of cells: numerical feature extraction methods, end-to-end neural network models, and transport-based morphometry and compares those approaches on four publicly available datasets.