Searcharxiv⌕ Search

arXiv subjects

Bijoy Kundu

Publications and source records attributed to Bijoy Kundu.

16 recordsLinked to original sources
arXiv2025

Machine Learning Derived Blood Input for Dynamic PET Images of Rat Heart

Dynamic FDG PET imaging study of n = 52 rats including 26 control Wistar-Kyoto (WKY) rats and 26 experimental spontaneously hypertensive rats (SHR) were performed using a Siemens microPET and Albira trimodal scanner longitudinally at 1, 2, 3, 5, 9, 12 and 18 months of age. A 15-parameter dual output model correcting for spill over contamination and partial volume effects with peak fitting cost functions was developed for simultaneous estimation of model corrected blood input function (MCIF) and kinetic rate constants for dynamic FDG PET images of rat heart in vivo. Major drawbacks of this model are its dependence on manual annotations for the Image Derived Input Function (IDIF) and manual determination of crucial model parameters to compute MCIF. To overcome these limitations, we performed semi-automated segmentation and then formulated a Long-Short-Term Memory (LSTM) cell network to train and predict MCIF in test data using a concatenation of IDIFs and myocardial inputs and compared them with reference-modeled MCIF. Thresholding along 2D plane slices with two thresholds, with T1 representing high-intensity myocardium, and T2 representing lower-intensity rings, was used to segment the area of the LV blood pool. The resultant IDIF and myocardial TACs were used to compute the corresponding reference (model) MCIF for all data sets. The segmented IDIF and the myocardium formed the input for the LSTM network. A k-fold cross validation structure with a 33:8:11 split and 5 folds was utilized to create the model and evaluate the performance of the LSTM network for all datasets. To overcome the sparseness of data as time steps increase, midpoint interpolation was utilized to increase the density of datapoints beyond time = 10 minutes. The model utilizing midpoint interpolation was able to achieve a 56.4% improvement over previous Mean Squared Error (MSE).

eess.IV↗
arXiv2025

Determination of unscaled blood input for human dynamic FDG brain PET

Objectives: Many existing techniques for the non-invasive quantification of the blood input function in dynamic FDG-PET imaging require strong historical information or user input. The technique proposed in this work utilizes the assumption that a dynamic PET scan can be modeled by the Patlak plot to determine an unscaled blood input function. Materials and Methods: The time activity curve (TAC) for each voxel in a dynamic image can be considered as an n-dimensional vector. In this context, a TAC follows the Patlak plot if and only if the TAC is a linear combination of the blood input function and the integral of the blood input function. Given a set of TACs which follow the Patlak plot, we can thus use PCA to determine a basis which spans the same vector space as the blood input function and the integral of the blood input function. We then seek to find two TACs in this vector space which best satisfy that the estimated anti-derivative of one of the TACs is close to the other TAC; such TACs are candidates for the blood input function and the integral of the blood input function. We were able to construct a low (2) dimensional optimization problem to find such TACs. Results: We applied our results to obtain predicted blood input functions and Ki maps for twelve normal subjects. Scaling the predicted blood input function to best match the ground truth, we achieved an average SSE of $0.042 \pm 0.032$ and an average DTW distance of $0.141 \pm 0.053$. Matching the means of the predicted and ground truth Ki maps, we achieved an average MAPE of $2.539 \pm 0.928$ and an average SSIM of $0.991 \pm 0.005$. Conclusion: While not often viewed as such, the assumption that some dynamic data follows a kinetic model gives strong prior information. In the case of the Patlak plot, we can use this assumption to estimate an unscaled blood input function and unscaled Ki map.

physics.med-ph↗
arXiv2024

An end-to-end deep learning pipeline to derive blood input with partial volume corrections for automated parametric brain PET mapping

Dynamic 2-[18F] fluoro-2-deoxy-D-glucose positron emission tomography (dFDG-PET) for human brain imaging has considerable clinical potential, yet its utilization remains limited. A key challenge in the quantitative analysis of dFDG-PET is characterizing a patient-specific blood input function, traditionally reliant on invasive arterial blood sampling. This research introduces a novel approach employing non-invasive deep learning model-based computations from the internal carotid arteries (ICA) with partial volume (PV) corrections, thereby eliminating the need for invasive arterial sampling. We present an end-to-end pipeline incorporating a 3D U-Net based ICA-net for ICA segmentation, alongside a Recurrent Neural Network (RNN) based MCIF-net for the derivation of a model-corrected blood input function (MCIF) with PV corrections. The developed 3D U-Net and RNN was trained and validated using a 5-fold cross-validation approach on 50 human brain FDG PET datasets. The ICA-net achieved an average Dice score of 82.18% and an Intersection over Union of 68.54% across all tested scans. Furthermore, the MCIF-net exhibited a minimal root mean squared error of 0.0052. The application of this pipeline to ground truth data for dFDG-PET brain scans resulted in the precise localization of seizure onset regions, which contributed to a successful clinical outcome, with the patient achieving a seizure-free state after treatment. These results underscore the efficacy of the ICA-net and MCIF-net deep learning pipeline in learning the ICA structure's distribution and automating MCIF computation with PV corrections. This advancement marks a significant leap in non-invasive neuroimaging.

eess.IV↗
arXiv2023

Multimodal Deep Learning to Differentiate Tumor Recurrence from Treatment Effect in Human Glioblastoma

Differentiating tumor progression (TP) from treatment-related necrosis (TN) is critical for clinical management decisions in glioblastoma (GBM). Dynamic FDG PET (dPET), an advance from traditional static FDG PET, may prove advantageous in clinical staging. dPET includes novel methods of a model-corrected blood input function that accounts for partial volume averaging to compute parametric maps that reveal kinetic information. In a preliminary study, a convolution neural network (CNN) was trained to predict classification accuracy between TP and TN for $35$ brain tumors from $26$ subjects in the PET-MR image space. 3D parametric PET Ki (from dPET), traditional static PET standardized uptake values (SUV), and also the brain tumor MR voxels formed the input for the CNN. The average test accuracy across all leave-one-out cross-validation iterations adjusting for class weights was $0.56$ using only the MR, $0.65$ using only the SUV, and $0.71$ using only the Ki voxels. Combining SUV and MR voxels increased the test accuracy to $0.62$. On the other hand, MR and Ki voxels increased the test accuracy to $0.74$. Thus, dPET features alone or with MR features in deep learning models would enhance prediction accuracy in differentiating TP vs TN in GBM.

eess.IV↗
arXiv2000

Progress in Perturbative Color Transparency

A brief overview of the status of color transparency experiments is presented. We report on the first complete calculations of color transparency within a perturbative QCD framework. We also comment on the underlying factorization method and assumptions. Detailed calculations show that the slope of the transparency ratio with $Q^2$, and the effective attenuation cross sections extracted from color transparency experiments depend on the $x$ distribuition of wave functions.

hep-ph↗
arXiv2000

Perturbative Color Transparency in Electroproduction Experiments

We calculate quasi-exclusive scattering of a virtual photon and a proton or pion in nuclear targets. This is the first complete calculation of ``color transparency" and "nuclear filtering " in perturbative QCD. The calculation includes full integrations over hard interaction kernels and distribution amplitudes in Feynman -x fractions and transverse spatial separation space $b$. Sudakov effects depending on $b$ and the momentum transfer $Q^2$ are included. Attenuation of the hadronic states propagating through the medium is calculated using an eikonal Glauber formalism. Nuclear correlations are included explicitly. We find that the color transparency ratio is comparatively insensitive to theoretical uncertainties inherent in perturbative formalism, such as choice of infrared cutoff scales. However, the $Q^2$ dependence of the transparency ratio is found to depend sensitively on the model of the distribution amplitude, with endpoint-dominated models failing to be dominated by short-distance. Color transparency experiments should provide an excellent test of the underlying theoretical assumptions used in the pQCD calculations.

hep-ph↗
arXiv2000

Oscillating Color Transparency in $πA\to πp (A-1)$ and $γA\to πN (A-1)$

The energy dependence of $90^o$ $cm$ fixed angle scattering of $πp \to π' p'$ and $γp\to π^+ n$ at large momentum transfer are found to be well described in terms of interfering short and long distance amplitudes with dynamical phases induced by Sudakov effects. We calculate the color transparency ratio for the corresponding processes in nuclear environments $πA\to π'p(A-1)$ and $γA\to πN (A-1)$ taking nuclear filtering into account. A prediction that the transparency ratio for these reactions will oscillate with energy provides an important test of the Sudakov phase shift and nuclear filtering hypothesis which is testable in upcoming experiments.

hep-ph↗
arXiv2000

Exploring vector meson masses in nuclear collisions

The formalism developed earlier by us for the propagation of a resonance in the nuclear medium in proton-nucleus collisions has been modified to the case of vector boson production in heavy-ion collisions. The first part of the talk describes this formalism. The formalism includes coherently the contribution to the observed di-lepton production from the decay of a vector boson inside as well as outside the nuclear medium. The calculated invariant rho mass distributions are presented for the $ρ$-meson production using optical potentials estimated within the VDM and the resonance model. In the second part of the talk we write a formalism for coherent rho production in proton nucleus collisions and explore the sensitivity of the (p,p$^\prime ρ^0$) reaction cross section to medium mass modification of the rho meson.

nucl-th↗
arXiv1999

Hadronic electromagnetic form factors and color transparency

We review the current status of electromagnetic form factor calculations in perturbative QCD. There is growing evidence that factorization prescriptions involving a transverse coordinate integration, such as that of Li and Sterman, are more appropriate than the prescription of LePage and Brodsky. Color transparency is naturally described within the formalism. We report the first explicit calculations of color transparency and nuclear filtering as perturbatively calculable phenomena.

hep-ph↗
arXiv1999

Resonance propagation in heavy-ion scattering

The formalism developed by Jain and Kundu for the propagation of a resonance in the nuclear medium has been modified to the case of heavy-ion collisions at relativistic energies. The formalism includes coherently the contribution to the observed di-lepton production from the decay of the resonance inside as well as outside the nuclear medium. The calculated results are presented for the $ρ$-meson production. It is observed that, in general, the shape, magnitude and peak position of the coherently summed invariant mass distribution is much different from that obtained by summing the inside and outside contributions incoherently. Therefore, while inferring the modification of hadron properties produced in heavy-ion collisions from experiments, it is important that in theoretical calculations one includes the decay of the propagating hadron from inside and outside the heavy-ion system coherently. We also find that the mass distribution is sensitive to the amount of the medium modification of the $ρ$-meson.

nucl-th↗
arXiv1999

New Results on Perturbative Color Transparency in Quasi-Exclusive Electroproduction

We review the perturbative QCD formalism of hadronic electromagnetic form factors and the color transparency ratio for quasi-exclusive electroproduction of the proton and pion from nuclear targets. We have completed the first full calculations including all leading order quark subprocesses and integrations over distribution amplitudes, including Sudakov effects. For the case of the proton, the calculated result shows scaling beyond $Q^2=10$ GeV$^2$. The calculation incorporating filtering due to the nuclear medium is cleaner than the corresponding calculation in free space because of attenuation of large distance amplitudes. We find that the color transparency ratio is rather insensitive to theoretical uncertainties inherent in the perturbative formalism, such as the choice of the hadron distribution amplitude.

hep-ph↗
arXiv1998

The elementary p(p,p'π^{+})n reaction

A detailed study of the elementary p(p,p$'π^{+}$)n reaction is presented using the delta isobar model. In this model, in the first step one of the two protons in the initial state gets excited to $Δ$. This, in the second step, decays into a nucleon and a pion. For the $pp \to NΔ$ step the parametrized form of the DWBA t-matrix of Jain and Santra, which reproduces most of the available data on $pp \to nΔ^{++}$, is used. The cross-sections studied include the outgoing proton momentum spectra in coincidence with the pion, the outgoing pion momentum spectra and the integrated total cross-section. We find that all the calculated numbers are in good agreement with the corresponding measured cross sections.

nucl-th↗
arXiv1998

The perturbative proton form factor reexamined

We recalculate the proton Dirac form factor based on the perturbative QCD factorization theorem which includes Sudakov suppression. The evolution scale of the proton wave functions and the infrared cutoffs for the Sudakov resummation are carefully chosen, such that the soft divergences from large coupling constants are diminished and perturbative QCD predictions are stablized. We find that the King-Sachrajda model for the proton wave function leads to results which are in better agreement with experimental data compared to the Chernyak-Zhitnitsky wave function.

hep-ph↗
arXiv1996

Delta Decay in Nuclear Medium

Proton-nucleus collisions, where the beam proton gets excited to the delta resonance and then decays to p$π^+$, either inside or outside the nuclear medium, are studied. Cross-sections for various kinematics for the (p,p$' π^+$) reaction between 500 MeV and 1 GeV beam energy are calculated to see the effects of the nuclear medium on the propagation and decay of the resonance. The cross-sections studied include proton energy spectra in coincidence with the pion, four momentum transfer distributions, and the invariant p$π^+$ mass distributions. We find that the effect of the nuclear medium on these cross-sections mainly reduces their magnitudes. Comparing these cross-sections with those considering the decay of the delta outside the nucleus only, we further find that at 500 MeV the two sets of cross-sections have large differences, while by 1 GeV the differences between them become much smaller.

nucl-th↗