Divya Saxena’s Post

I am glad to share our latest paper "DuDoCFNet: Dual-Domain Coarse-to-Fine Progressive Network for Simultaneous Denoising, Limited-View Reconstruction, and Attenuation Correction of Cardiac SPECT" published on IEEE Transactions on Medical Imaging (TMI). In this work, we proposed a cross-modal and cross-domain framework for simultaneous denoising, attenuation correction, and limited-view reconstruction of cardiac SPECT. Our learning strategy demonstrated superior accuracy in image prediction and reconstruction compared to existing single- or multi-task learning methods, across various low-dose levels. Paper link (open access): https://lnkd.in/exBhmzzA Code link (open source): https://lnkd.in/ezrfb4mv Many thanks to my advisor Chi Liu, collaborators Bo Zhou, Xueqi Guo, Huidong Xie, Qiong L., James Duncan, Sinusas Albert, reviewers, and editors. Please reach me if you have any questions :) #deeplearningai #medicalimaging #spectct #reconstruction #denoising #attenuationcorrection

Hyperspectral imaging is revolutionizing fields like ocean monitoring and medical diagnostics. The ELO-Hyp Collaborative Research Project, funded under the #NorwayGrants and coordinated by Prof. Necoară, focuses on developing advanced learning and optimization algorithms to enhance imaging models. 📈✨ 🔍 Key objectives included: ✅ Creating and benchmarking algorithms using real datasets; ✅ Ensuring interoperability across imaging platforms; ✅ Producing free software for broader accessibility. Watch this episode to discover how ELO-Hyp is set to transform our understanding of complex environments and medical applications! ___ #ResearchInTheSpotlight is a video podcast for scientists and passionate science people having as protagonists the principal investigators (PI) of the Collaborative Research Projects funded in 2018 & 2019 calls under the Romanian Research Programme - EEA and Norway Grants. This video podcast is designed by the #UEFISCDI, as Programme Operator of the Research Programme funded under the EEA and Norway Grants. ⏬ University POLITEHNICA of Bucharest Norwegian University of Science and Technology (NTNU) University of Bucharest Spitalul Clinic Colțea https://lnkd.in/dyP6GQ9h

I am pleased to announce the publication of our research paper "Classifying and Quantifying MRI Image Quality from DICOM Data at the Edge" - in the 2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE). Our paper addresses a key issue in radiology: improving MRI image quality while reducing scan duration, which is essential for both diagnostic accuracy and patient comfort. We introduce "MriqNet - a Deep Learning-based framework" that classifies MRI images and quantifies image quality using metadata from DICOM images. Through integration with our open-source framework, Niffler, we successfully implemented MriqNet in cloud and edge environments, demonstrating its effectiveness in automating image quality assessment. This work would not have been possible without the mentorship and guidance of Professor Pradeeban Kathiravelu, Ph.D., whose continued expertise and support were invaluable throughout the research process. I am deeply grateful for the opportunity to work under his guidance and leadership. I would also like to express my gratitude to Rochan singh, whose collaboration and dedication contributed significantly to the success of this project. We look forward to further advancing the applications of AI in medical imaging and contributing to the ongoing efforts in enhancing diagnostic capabilities in healthcare. Please find the link to the research paper attached to this post - https://lnkd.in/ghvxzb6c #research #AI #radiology #deepLearning #healthcare #DICOM #MriqNet #IEEE #innovation #academicresearch

❓ Can Deep-learning denoising minimize radiation exposure 📉in neck CT beyond the limits of conventional reconstruction? 💡 Honored to present you the #EJR paper of the week 🖇️ https://lnkd.in/dhQFRqaQ   🇪 🇸 🇸 🇪 🇳 🇹 🇮 🇦 🇱 🇸 1️⃣ In conventional reconstruction techniques, limiting radiation dose comes at the cost of diminished diagnostic accuracy. 2️⃣ AI may be of help to solve this dilemma. 3️⃣ This paper evaluated an AI-based denoising post-processing software in neck CT. 4️⃣ This approach enabled diagnostic-quality neck CT images with radiation doses reduced to 25 % of conventional levels, thus significantly minimizing patient exposure. 🤝 congratulations to the authors: @David Plajer, Marlene Hahn, @Marianna Chaika, @Markus Mader, @Jonas Mueck, Prof. Dr. Konstantin Nikolaou, PD Dr. Saif Afat, Assoc. Prof. (PD) Dr. med. Andreas S. Brendlin   #RadiologyAI #LowDoseImaging #NeckCT #DeepLearningDenoising #PatientSafety Deutsche Röntgengesellschaft e.V. Siemens Healthineers

🧠 Trivia Time! In the context of digital twins in medical simulations, which core challenge must be addressed to enable real-time patient-specific simulations for personalized medicine? 🤖🏥 A) High-resolution imaging for accurate anatomical modeling B) Integration of heterogeneous data sources for comprehensive modeling C) Sufficient computational power and algorithms for real-time processing D) Advanced visualization techniques for user-friendly interface #DigitalTwins #MedicalInnovation #PersonalizedMedicine #HealthcareTech Sciqst, Precise References in Medicine. www.sciqst.com

The journey of medical imaging, from the early days of the microscope to the advanced realms of MRIs and beyond, is a testament to the relentless pursuit of scientific advancement. In "The Past, Present, and Future of Medical Imaging," the group at CZI delve into the milestones defining this fascinating field, showcasing how each leap forward has offered unprecedented insights into the imaging technology field. As we look to the future, KUBTEC® remains dedicated to pushing the boundaries of what's possible in X-ray imaging systems, ensuring that the next generation of scientific researchers has the tools to see clearly, and research efficiently. Read the full article here: https://lnkd.in/eNY_2AVJ #Xrayimaging #Innovation #KUBTEC #HealthcareTechnology #FutureOfMedicine #ScientificInnovation #MedicalTechnology #LabEquipment #ResearchTools #BiotechSolutions #HealthcareInnovation #DigitalPathology #ImagingTechnology #PrecisionMedicine #LifeSciences #MedTechAdvancement #LabScience #HealthTech #DiagnosticImaging #InnovativeSolutions #KubtecScientific #KubtecImaging #KubtecLabSolutions #KubtecPrecisionInstruments #NonDestructiveTesting #XrayImaging

Just one a week left until this year's BVM. In addition to the scientists already introduced, PhD student Saikat Roy from our group will also be present and giving an oral presentation. On Tuesday, March 12th, he will speak at 10am on "MedNeXt: Transformer-Driven Scaling of ConvNets for Medical Image Segmentation". His talk is based on his identical paper where Saikat discusses the growing interest in using #Transformer -based architectures for medical #image #segmentation. However, due to the scarcity of large annotated medical datasets, achieving comparable performance to natural images is challenging. Convolutional networks, with their higher inductive biases, are more easily trainable to high performance. The paper introduces MedNeXt, a modernized convolutional architecture tailored to data-scarce medical settings. MedNeXt includes features such as a fully ConvNeXt 3D Encoder-Decoder Network, residual ConvNeXt up and downsampling blocks, a technique to iteratively increase kernel sizes, and compound scaling. These innovations result in state-of-the-art performance on CT and MRI modalities across various dataset sizes, making MedNeXt a promising deep architecture for medical image segmentation. The work has garnered popular attention, with nearly 200 stars on GitHub, and is made publicly available at: https://lnkd.in/e9pZt-KQ. And here you can access the paper: https://lnkd.in/eMuEm8R7.

Taking a #differentPerspective: Yesterday, I attend the Computing in Cardiology 2024 (#CinC2024) conference, hosted at Karlsruher Institut für Technologie (KIT) and found many common research questions between #industrialAI and #cardiacComputation: ▶ How to foster #trustworthiness in AI-based systems (e.g. with #explainableAI and #mlOps) ▶ How to improve #stability and #speed of simulations as well as #inverseProblems (e.g. with #PINNs or #GNNs) ▶ How to use AI to overcome the #mediaBreaks (e.g. using #DNNs for extracting information from paper prints) ▶ How will the #EUaiAct impact the product life-cycle? Maybe these are some leads for potential collaboration? Approach me in case of interest to discuss! And of course it was great to meet old friends again 😊 Blanca Rodriguez, Jichao Zhao, Axel Loewe, Pablo Lamata, Olaf Doessel, Pyotr Platonov, Fernando Campos, Dr. Gunnar Seemann, Martin Bishop and to discuss with current colleagues and collaborators Leonie Schicketanz and Jakub Grzelak! Computing in Cardiology #modelling #simulation #signalProcessing #AI #physicsInformedNeuralNetworks #graphNeuralNetworks #aiAct

📹 Check out our latest video where we dive into the amazing capabilities of the MI-SIM. We compare traditional imaging results with those achieved using our intelligent tracking feature. The improvements with MI-SIM are amazing! Be sure to like, comment or share and enjoy your weekend with a glimpse into the future of scientific imaging! #Microscopy #Innovation #ScienceTechnology #CSRbiotech #MISIM

Quantitative phase imaging (QPI) techniques offer intrinsic information about #biomedical samples in a noninvasive, label-free way. Learn about QPI and its enormous potential for applications, in this #OpenAccess review published in #SPIE_JBO. https://lnkd.in/gSdcUt_W