https://lnkd.in/ewZ_kTik https://lnkd.in/eEV93dev https://lnkd.in/e5njE_2T 🚀 Excited to share three projects I have been working on as part of my journey into Machine Learning and Medical Imaging! 🔬 Project 1: Kidney CT Scan Abnormality Detection A dual-model pipeline combining YOLOv11 (object detection) and YOLOv8 (classification) to detect Cyst, Stone, Tumor, and Normal kidneys from CT scans — achieving 100% Top-1 accuracy on the classification task. 📎 GitHub: https://lnkd.in/eEV93dev 📎 Kaggle Notebook: https://lnkd.in/eXSxRkRB 📈 Project 2: Time Series Analysis — California Births Exploring temporal patterns and forecasting using classical time series techniques. 📎 GitHub: https://lnkd.in/e5njE_2T I am actively learning and growing in this field, and I would love to hear from anyone with feedback, suggestions, or recommendations on how I can improve these projects. Any guidance from experienced practitioners, researchers, or clinicians is truly welcome! 🙏 #MachineLearning #DeepLearning #MedicalImaging #YOLO #ComputerVision #HealthcareAI #DataScience #OpenToFeedback

New Post: Adaptive Multi‑Scale CLAHE–CNN for Functional MRI Micro‑vasc­ular Enhancement - **Abstract** \(≈235 words\) Enhancing micro‑vascular contrast in functional magnetic resonance imaging \(fMRI\) is essential for accurate detection of subtle neural perfusion changes. Conventional contrast‑limited adaptive histogram equalization \(CLAHE\) improves global visibility but introduces spatially inconsistent sharpening and noise amplification, particularly in high‑frequency cortical structures. We propose **Adaptive Multi‑Scale CLAHE–CNN \(AMS‑CNN\)**, a hybrid framework that integrates \[…\] \[Source & Legal Disclaimer\] This is an AI-generated simulation research dataset provided by Freederia.com, released under the Apache 2.0 License. Users may freely modify and commercially use this data \(including patenting novel improvements\); however, obtaining exclusive patent rights on the original raw data itself is prohibited. As this is AI-simulated data, users are strictly responsible for independently verifying existing copyrights and patents before use. The provider assumes no legal liability. For future Enterprise API access and bulk dataset purchase inquiries, please contact Freederia.com.

Mantis Biotech Deploys Physics-Based Digital Twins to Combat Medical Data Scarcity 📌 Mantis Biotech is revolutionizing medical research by creating physics-based digital twins of the human body - synthetic models that simulate rare diseases and complex surgeries using AI, motion data, and biomechanics. This breakthrough bypasses real-world data scarcity, enabling safer training, better drug testing, and smarter injury prevention - all without compromising patient privacy. 🔗 Read more: https://lnkd.in/dHyeHe39 #Mantisbiotech #Physicsengine #Digitaltwins #Llms #Medicalsimulation

𝐈𝐧𝐭𝐫𝐨𝐝𝐮𝐜𝐢𝐧𝐠 𝐌𝐑-𝐑𝐀𝐓𝐄: 𝐚 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐟𝐨𝐫 𝐛𝐫𝐚𝐢𝐧 𝐌𝐑𝐈 𝐢𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 We’re excited to introduce MR-RATE, a large-scale dataset and vision-language foundation model for brain MRI, developed by Forithmus in collaboration with NVIDIA and made accessible to the research community. MR-RATE includes over 𝟏𝟎𝟎,𝟎𝟎𝟎 𝐛𝐫𝐚𝐢𝐧 𝐌𝐑𝐈 𝐬𝐭𝐮𝐝𝐢𝐞𝐬 from more than 𝟖𝟎,𝟎𝟎𝟎 𝐮𝐧𝐢𝐪𝐮𝐞 𝐩𝐚𝐭𝐢𝐞𝐧𝐭𝐬, with ~𝟕𝟎𝟎,𝟎𝟎𝟎 𝟑𝐃 𝐯𝐨𝐥𝐮𝐦𝐞𝐬 paired with de-identified 𝐫𝐚𝐝𝐢𝐨𝐥𝐨𝐠𝐲 𝐫𝐞𝐩𝐨𝐫𝐭𝐬 𝐚𝐧𝐝 𝐫𝐢𝐜𝐡 𝐜𝐥𝐢𝐧𝐢𝐜𝐚𝐥 𝐦𝐞𝐭𝐚𝐝𝐚𝐭𝐚, reflecting real-world neuroimaging practice (all available on Hugging Face). On top of this, we are developing vision-language models that jointly understand 3D MRIs and generate clinically meaningful radiology reports, supporting radiologists by automating time-intensive parts of interpretation while keeping clinicians in the loop. With MR-RATE, we continue moving beyond narrow AI tools toward generalist medical intelligence systems. Our goal is simple: make medical intelligence accessible everywhere. This will happen through open and collaborative research. Code: https://lnkd.in/eU3NWEiM Dataset: https://lnkd.in/ed6x_3rt Read more: https://lnkd.in/eaSx8ZzN

It has been great to work with Forithmus, UZH, and Medipol University Hospital on the creation and publication of the MR-RATE dataset. NVIDIA used MR-RATE to train the NV-Generate-MR-Brain model available with an NVIDIA Open Model License. https://lnkd.in/eSha43mc.

Really excited to share this—NVIDIA has been building MR-RATE dataset with Forithmus, University of Zurich. Building foundation models for medical imaging is challenging, especially when dealing with 3D data and real-world clinical variability. MR-RATE is a big step toward solving that by bringing scale, quality, and multimodal alignment together. This is about more than a dataset or a model—it’s about enabling a new generation of medical AI that can actually integrate into clinical workflows. Grateful to be part of this collaboration and excited for what comes next.

HLIP now supports the MR-RATE dataset. In a very preliminary experiment, we achieved an AUC of 76.4% on zero-shot evaluation over the test split. Training takes 16 hours on 8 L40 GPUs. The HLIP model (https://lnkd.in/grC-wfmS) trained on our internal dataset achieves an AUC of 80.3% on zero-shot evaluation over the MR-RATE test split. Code: https://lnkd.in/gN8-uHbd

See how Hydra Bio Plasma FIB imaging and Amira Software reveal brain membranes, vesicles, mitochondria & myelin in stunning detail. Amira’s AI-powered tools streamline segmentation, highlighting the intricate complexity of neural cells. 🧠✨ https://bit.ly/4bO5uYL #AmiraSoftware #FIBSEM

See how Hydra Bio Plasma FIB imaging and Amira Software reveal brain membranes, vesicles, mitochondria & myelin in stunning detail. Amira’s AI-powered tools streamline segmentation, highlighting the intricate complexity of neural cells. 🧠✨ https://bit.ly/4c7T4ei #AmiraSoftware #FIBSEM

Diagnostic Imaging using Machine Learning Imaging may be integrated to your process control to activate actions or perform diagnostics. Images can be created using Xrays, CTscans, Thermal Imaging, you name it. https://lnkd.in/gB4GPiqi