Verily’s Post

The aorta is our body’s primary artery and is responsible for delivering oxygen-rich blood throughout our system. Why should you care about this and what does it have to do with Rackspace Technology? The University of Galway is on the brink of innovative medical solutions to reduce the threat of aortic disease. Through computer simulations tailored to individual patients, the Virtual Aorta initiative is clearing the path for more timely and accurate diagnoses of aortic complications. However, this kind of breakthrough requires real-time aortic diagnosis - something that is usually hampered by the intricacies of modelling blood flow. But, with the backing of Rackspace Technology and AWS’s cloud infrastructure, the constraint of relying on local computing resources was lifted. We reduced the computational modelling process from 150 hours to 10 minutes with public cloud scalability and AI/ML services. Now, with the transformative power of cloud services facilitated by the OCRE grant, the research team can direct its energies toward pivotal research outputs. Read the full case study here: https://lnkd.in/gDKaR2wn I’m also proud to say that we won the  'Education Partner of the Year - EMEA' award at the Partner Awards Gala 2023 during AWS re:Invent 2023 for this case study. And DM me if you want to achieve more with your AWS platform. #CloudComputing #DigitalTransformation #Data

I am pleased to have had the opportunity to present our most recent work, "SegmentOR: Obtaining Efficient Operating Room Semantics Through Temporal Propagation," at #MICCAI2023 in Vancouver yesterday! Automated point cloud segmentation provides a nuanced understanding of the surgical environment, which is fundamental for improving surgical procedures' efficiency and safety through intelligent systems and data-driven analysis. - **SegmentOR** introduces a method for weakly-supervised semantic segmentation in surgical operating rooms, utilizing 4D point cloud sequences. - It operates with less than 0.01% of annotated data, drastically lowering the burden of annotating surgical scenes. - The method achieves a segmentation mIoU 10% above baselines, utilizing fewer labels. Many thanks to my co-authors Daniel Derkacz-Bogner, Tony Wang, Benjamin Busam, and Nassir Navab. For a deeper dive into our work, check out our project page: https://lnkd.in/e_7CtyY3 And if you're around MICCAI the next few days and want to chat about Surgical Data Science, don't hesitate to reach out! #SurgicalDataScience #SurgicalSceneUnderstanding #ContextAwareSystems

🌍✨ Discover the conference publication entitled "Performance Experiences From Running An E-health Inference Process As FaaS Across Diverse Clusters." Read it here: https://lnkd.in/drjthA9M

REMINDER: We'll be attending the Bio-IT World next week! 📅 𝗪𝗵𝗲𝗻: Apr 15-17 ✈ 𝗪𝗵𝗲𝗿𝗲: Boston, MA, USA 👨💼 𝗪𝗵𝗼: Alan Lorimer & John Waller ❔ 𝗪𝗵𝘆: A detailed understanding of a laboratory/analytical workflow, software APIs, and database configuration is critical to fully exploiting cloud computing's storage and compute power. Our cloud solution, QDPconnect, gives scientists flexible control of cloud resources (CPU/GPU number and size) to power image analysis software for Digital Pathology yet easily applies to other analysis modalities. We'll be discussing our solutions and networking with peers across the various life science fields at Bio-IT. John will also be presenting our scientific poster, 'Optimising Workflows and Enhancing Productivity of Scientists through Cloud Computing Solutions'. We look forward to chatting with you more about Sciento, QDPconnect, our poster, and more. See you in Boston! —— #qdpconnect #bioitworld #bioitexpo #cloudcomputing #digitalpathology

If you’ve ever wondered about the working life of a young scientist involved in research in aging, check out our latest blog post and learn about “the wonderfully dichotomous, purple compound azure blue.” https://lnkd.in/gQaejmvE

Time to use the largest instance for single cell atlases and stream everything in compute memory! 😊😂😃 Nice to see this programmatic exploration and retrieval of the harmonised, curated and reannotated CELLxGENE single-cell human cell atlas now available in R. https://lnkd.in/dQ_uWcmu

More than 6,000 Rare Diseases have been identified.  See below for how OM1 can help. AI has the power to fill in gaps to identify patterns in databases. This core process can be used to identify mis- or underdiagnosed rare disease patients faster, shortening the time to diagnosis. Identifying patients in real-world data contexts can be difficult, especially when clinical coding is imprecise or unavailable. With OM1 PhenOM™, alongside our Real-World Data Cloud, OM1 can provide powerful solutions to rare disease challenges. Interested in learning more about how OM1 improves research efficiency, finds under-diagnosed patients, and supports clinical decision-making in rare diseases.? Check out these links: PhenOM ebook: 📍 https://lnkd.in/dA8HrjAN Book a Meeting: 📍 https://lnkd.in/eJcKBAJK #RareDisease #ArtificialIntelligence #PatientIdentification 

The dust has settled, and somehow RSNA 2023 is in the books. This was by far our most successful conference to date, and we were energized by the reception we received. Some highlights from us:   A week ago tonight, we cohosted an intimate event with our partners at Centaur Labs for our customers and friends. This was an opportunity to say thank you to our valued community with some great food, drinks, and company. If you attended and helped to make it so successful, thank you!   We also thought we'd take a moment to share some other insights and developments that we saw from the world of radiology AI that are shaping our future.   AI: The New Mainstream in Radiology. AI's integration into daily practices isn't just a pipedream anymore, it is accelerating and is being used to enhance diagnostic precision and patient care.   Foundational Models: The First Movers Hit the Market. The opening day of RSNA saw HOPPR launch "Grace", a multi-modal foundation model for medical imaging, with huge potential! We are proud to have played a role in providing the staggering quantity of data needed to train Grace (over a petabyte). As our our CEO Joshua Miller Miller wrote about in a recent opinion piece, we've been seeing more and more requests for foundational model-level datasets.    The Persistent Challenge: Access to Large Datasets. Despite the advancements, one challenge remains: the need for quick and efficient access to large, unbiased datasets. The richness of data is crucial for the effectiveness of AI, yet accessing these vast datasets remains a hurdle for many in the field (spoiler, this is where we step in). Early Access to Our Subscription Service: An Invitation. Following promising feedback from our poll in last month's newsletter, we are launching early access to our subscription service. This initiative is designed to empower developers with direct access to our tools and datasets, streamlining workflows and expanding capabilities while giving you economies of scale. We 'soft launched' this model with a special offer at RSNA and we were so encouraged by the uptake that we thought we'd extend it. So if instant access to huge datasets via a subscription model sounds like it makes sense for you, book a meeting with our team and enjoy some early adopter savings! Until next time McCormick Place! #medicalimaging #rsna2023 #aiinmedicine | Ouwen Huang, PhD | Rod B. Greco | Lucas Alves

FlexView Fun Fact Friday: Did you know that FlexView has features designed specifically for #DigitalPathology?! The Microscopy Mode in FlexView is equipped with a set of tools for annotation and defining regions of interest that are relevant to a digital pathology workflow, improving your focus. AND you can load these large files from the cloud at lightning speed in FlexView. Learn more here https://lnkd.in/eGKtxj33 #digitalmicroscopy #pathology #microscopy #pathologyinnovation #medicalimaging #flexview

Imagine seeing the tiny machines inside a cell!  That's what cryo-electron tomography (cryoET) does, but the images are like giant puzzles we need to label. Now, with AI tools like MemBrain, we can annotate these images super fast, like finishing a year's worth of work in just a few days!  This is all thanks to the cryoET Data Portal, where scientists share their images and collaborate. The future is bright! We're going to make amazing discoveries about how cells work, and that means better understanding of health, disease, and even how we age! Paper: https://lnkd.in/gAHRhN28