Catarina Moreira’s Post

🎉 Exciting Milestone Achieved! 🎉 I’m excited to announce, a bit later than planned, that our first co-authored paper has been published in the Journal of the American Society For Mass Spectrometry (ASMS), under the American Chemical Society. 🔬 Our tutorial paper provides a comprehensive guide on applying Random Forest algorithms to improve the analysis of Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) imaging data. As I love mosaics, I sometimes imagine MSI features as molecular tesserae--the small, individual pieces that make up such an artwork -- each carrying unique information. However, when information is fragmented across multiple tiles, it can lead to redundancy and complexity, making interpretation challenging. This fragmentation, caused by collinearity in MSI data, can obscure the true molecular patterns we aim to understand. In our approach, we recombine these fragmented pieces, blending similar information to clarify the molecular mosaic. This not only simplifies the computational models, making them more efficient and generalizable, but also enhances the clarity and intuitiveness of our interpretations. Ultimately, this leads to a deeper and more accessible understanding of the molecular structures we study. 🙏 I’m grateful to be part of a driven team and extend my heartfelt thanks to my co-authors Mariya Shamraeva, Stefanos Zoidis, Ian Anthony, and Sebastiaan Van Nuffel for their guidance and collaboration. Your support and hard work made this achievement possible! 🌟 This is just the beginning! 🌟 I’m excited to continue exploring the intersection of machine learning and mass spectrometry imaging, where each discovery adds a new block to our expanding molecular mosaic. 📄Read our paper and.... 💬Join the conversation: I'd love to hear your thoughts on our findings. #M4i #MassSpectrometry #MachineLearning #Imaging #MaastrichtUniversity

🎉 **Exciting News to Share!** 🎉 I am thrilled to announce that our paper, "MetaV: A Pioneer in Feature Augmented Meta-Learning Based Vision Transformer for Medical Image Classification," has been accepted for publication in the esteemed journal, Interdisciplinary Sciences: Computational Life Sciences, Publisher: Springer (IF:4.8). This work represents a significant step forward in the field of medical image analysis, introducing a novel approach that combines the power of meta-learning with the innovative architecture of Vision Transformers. Our findings demonstrate the potential to enhance diagnostic accuracy and offer new insights into disease identification and progression. I would like to extend my deepest gratitude to my co-authors Dr. Arun Prakash Agrawal Dr. Mohd Anas Wajid Mohd. Saif Wajid Aasim Zafar for their invaluable contributions and support throughout this journey. This achievement reflects our shared commitment to advancing research that can make a real difference in the lives of patients around the world. Stay tuned for the publication link and further updates. Your thoughts, feedback, and questions are always welcome! #MedicalImaging #MachineLearning #VisionTransformers #MetaLearning #AIinHealthcare #ResearchPublication

I am glad to share that our research paper, "Classification of High-Resolution Chest CT Scan Images Using Adaptive Fourier Neural Operators for COVID-19 Diagnosis" has been published in COVID as part of the Special Issue Artificial Intelligence and Machine Learning Applications for Developing the Diagnosis of COVID-19 and is available online. This research paper is fully open access and anyone can freely access this paper. COVID is a peer reviewed open access journal which is fully dedicated to research related to COVID-19 and Indexed in Emerging Sources Citation Index(ESCI), SCOPUS, OSTI (U.S. Department of Energy) etc. I highly appreciate all the authors of this paper for their valuable contribution. The publication link is as follows, https://lnkd.in/ePU_BZ2h #ESCI #Scopus #COVID #Research

Alhamdulillah, we are glad to share that our paper "Collaborative Learning for Annotation-Efficient Volumetric MR Image Segmentation" has been accepted and published online in Journal of Magnetic Resonance Imaging (JMRI). This research tackles the challenge of segmenting 3D MR images with limited labeling data (only one 2D slice per image). It proposes a novel method combining semi-supervised and self-supervised learning to generate reliable "pseudo labels" for unlabeled slices. This boosts segmentation accuracy compared to existing methods, making it promising for real-world scenarios where extensive manual labeling is impractical. https://lnkd.in/g6AwN6ng

⚡ New #causalML paper at #NeurIPS: 𝐃𝐢𝐟𝐟𝐏𝐎: 𝐀 𝐜𝐚𝐮𝐬𝐚𝐥 𝐝𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥 𝐟𝐨𝐫 𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐝𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧𝐬 𝐨𝐟 𝐩𝐨𝐭𝐞𝐧𝐭𝐢𝐚𝐥 𝐨𝐮𝐭𝐜𝐨𝐦𝐞𝐬 👉 We predict the potential outcomes of #treatments (e.g., in medicine or marketing). This is relevant where you don't want to simply understand the relative benefit of one treatment but where you want precise estimates of the outcomes themselves (e.g., for risk assessment). 🧐 Our model allows for rigorous #uncertainty #quantification (=predictive intervals). This is relevant for many decision-making tasks in medicine. 💪 We develop a novel #diffusion #model for this purpose. 💡 Our diffusion model has an *orthogonal* loss which makes it robust against model misspecification. https://lnkd.in/dJjrfpdQ with Yuchen M. Valentyn Melnychuk Jonas Schweisthal Munich Center for Machine Learning relAI Konrad Zuse School of Excellence

📢 #HighlyCited 1. Dual Band Antenna Design and Prediction of Resonance Frequency Using Machine Learning Approaches https://lnkd.in/g275mA9T 2. Detecting Malignant Leukemia Cells Using Microscopic Blood Smear Images: A Deep Learning Approach https://lnkd.in/gX3fJD6U 3. Combining LoRaWAN and NB-IoT for Edge-to-Cloud Low Power Connectivity Leveraging on Fog Computing https://lnkd.in/gBz4RBj5 4. 5G Based on MNOs for Critical Railway Signalling Services: Future Railway Mobile Communication System https://lnkd.in/gbdCNsDR MDPI Encyclopedia MDPI

Looking forward to presenting our work on "Determinstic global optimization for sample-efficient molecular design with generative ML" at the #NeurIPS2024 #AI4Mat workshop next week in Vancouver. We propose deterministic global optimization of molecular property prediction models in the form of ANNs trained on continuous latent spaces of generative ML models, such as VAEs. Using ANNs with ReLU activations, we formulate molecular design as a mixed-integer linear program (MILP) guaranteeing optimal molecular properties, as predicted by the ANN. Our results show superiority of the identified molecules with globally optimal predicted property values compared to those found with frequently used optimization strategies such as Bayesian optimization, indicating the potential to increase sample efficiency in ML-based molecular discovery workflows. For more details, you can find the conference paper at: https://lnkd.in/eaZ9bKiV Many thanks to Malte Franke and Alexander Mitsos for this great collaboration. #machinelearning #moleculardesign #CAMD #DGO #MILP

🚀 Exciting News! 🚀 I am thrilled to share that our latest research paper, "Frequency and Uncertainty driven Deep Learning Approach to Segment Electrocardiogram Signals for Effective Heart Parameters Estimation," has been presented at the prestigious IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI24)! This paper is the result of an incredible collaboration with Gennaro Esposito Mocerino, Muhammad Salman Haleem, Leandro Pecchia, and Alessio Botta. In this study, we tackled one of the biggest challenges in ECG analysis: accurate beat-level segmentation for reliable heart rhythm classification. We designed a multi-layer deep learning model to classify ECG signals into PQ, QRST, and TP intervals. By leveraging PhysioNet’s QT database, our model integrates unique features like short-time Fourier transform and approximate entropy to enhance classification with uncertainty-informed results. 💪🔬 Our findings pave the way for better heart rhythm parameter detection and hold exciting potential for the future of ECG signal processing. Huge thanks to this dedicated team and everyone who supported this work! 🔗 Read the full paper here: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://lnkd.in/duf5TVVQ #Research #DeepLearning #ECG #SignalProcessing #HealthcareInnovation #MachineLearning

🌟 Exciting News Alert! 🌟 Thrilled to share that our latest research paper has been accepted in the prestigious IEEE Reviews on Biomedical Engineering (IEEE RBME)!🎉✨ This article represents a collaborative journey with amazing partners from Stanford, Monash, Munich, Hong Kong University of Science and Technology, and more! 🌐 Together, we’ve explored groundbreaking ways to accelerate MRI scanning times, enhancing patient comfort and image quality by harnessing the power of data and physics-driven models! 📈💡 🔗 Check out the full paper here: https://lnkd.in/d8RqPSkE Here's a glimpse into what we’ve covered: 🧠 The landscape of deep learning in MRI reconstruction: from algorithm unrolling, Plug-and-Play, enhancement-based methods, to the future of generative models like VAEs, GANs, Diffusion Models, and more! ⚙️ Integrating data models with physics insights: pioneering advancements in hardware accelerations and sampling optimization to improve imaging. 💡 Tackling domain-specific challenges, from data scarcity and evaluation metrics to improving model generalization and memory constraints. 🌍 This work highlights the path forward for MRI in clinical settings, aiding researchers and clinicians to adopt the latest advancements in MRI technology. Grateful to our fantastic co-authors and collaborators, and especially to everyone who supported us along the way! 🙏 Looking forward to sharing the impact of this work with the research and clinical communities! 💬 Please let us know your comments, and we will make it a continuous update via https://lnkd.in/d5df7n9F in this field. #MRI #BiomedicalEngineering #IEEE #DeepLearning #GenerativeModels #MedicalImaging #DataScience #PhysicsBasedModeling #Research #Collaboration #HealthcareInnovation

📃Scientific paper: Neuromorphic Synergy for Video Binarization Abstract: Bimodal objects, such as the checkerboard pattern used in camera calibration, markers for object tracking, and text on road signs, to name a few, are prevalent in our daily lives and serve as a visual form to embed information that can be easily recognized by vision systems. While binarization from intensity images is crucial for extracting the embedded information in the bimodal objects, few previous works consider the task of binarization of blurry images due to the relative motion between the vision sensor and the environment. The blurry images can result in a loss in the binarization quality and thus degrade the downstream applications where the vision system is in motion. Recently, neuromorphic cameras offer new capabilities for alleviating motion blur, but it is non-trivial to first deblur and then binarize the images in a real-time manner. In this work, we propose an event-based binary reconstruction method that leverages the prior knowledge of the bimodal target's properties to perform inference independently in both event space and image space and merge the results from both domains to generate a sharp binary image. We also develop an efficient integration method to propagate this binary image to high frame rate binary video. Finally, we develop a novel method to naturally fuse events and images for unsupervised threshold identification. The proposed method is evaluated in publicly available and our collected data sequence, and shows the proposed method can ... Continued on ES/IODE ➡️ https://etcse.fr/vmCLq ------- If you find this interesting, feel free to follow, comment and share. We need your help to enhance our visibility, so that our platform continues to serve you.