🎉 Our PhD student @1richardobinna presents at #MICCAI2024! His paper on multi-conditional latent diffusion models could revolutionise DCE-MRI without the need for contrast agents. Catch his poster today! Full paper: https://t.co/3D9iMxh8aF 💡🧠 #AIinHealthcare #MedicalImaging

🏆 This year's First Place Paper Award at the EMERGE workshop went to Mehmet Yiğit Avcı, Emily Chan, @cosmin_bercea et al. for their work on Unsupervised Analysis of Alzheimer’s Disease Signatures! Congratulations from all the organisers! 🔗 https://t.co/vMhkj50jPx #MICCAI2024

🤖 Can we tailor the reverse process of diffusion models for better diagnostics? Discover THOR at #MICCAI 24, improving anomaly detection in brain MRIs 🧠 and wrist X-rays🖐️. Pre-print: https://t.co/NyZKj85skd Code: https://t.co/IgbAk1JSi6 🙏Bene @DanielRueckert @ja_schnabel

📢 PhD/PostDoc Opportunity Fully-funded (healthcare, pension, >50T€ pa) position in my group @uni_bamberg_of Opportunity for impactful ML/CV research and to publish & attend prestigious conferences (e.g., CVPR/ICCV/MICCAI). Topic flexible. Details below. Please share/RT🙏

What do autoencoders learn? Challenging common assumptions in anomaly detection by Cosmin Bercea at #MICCAI2023

@cosmin_bercea presenting his second paper on anomaly detection at #MICCAI2023

https://t.co/UdiJOUSdnm

4/N, N=4: We propose MorphAEus, deformable AEs. The AE focuses on identifying features of the healthy distribution, while the deformation component refines sharp boundaries 🎯. Dive in 🔗: Project: https://t.co/kQVVePbHtk Code: https://t.co/JKCiSViKJe

3/N: SOTA AEs, despite their advancements, often grapple with the same challenges. Advanced adversarial AEs, exhibit robust generative capabilities but may create input-reconstruction pairs that mismatch, another door to false positives❌.

2/N: While AEs trained on healthy samples 💚 are ideally expected to proficiently reconstruct only these samples, real-world results beg to differ. Shallow AEs tend to copy everything 🔄, anomalies included. Deep AEs yield blurry outputs, leading to false positives ❌

Can Autoencoders (AEs) really capture what's crucial for #AnomalyDetection? 🕵️‍♀️Dive into how unsupervised representation learning combines with image registration to excel in anomaly detection! 🦠🔍 Join us in Vancouver! 🤝 #MICCAI2023 Thread 1/N 🧵 @DanielRueckert @ja_schnabel

📢 From #MICCAI2023 #FAIMI: Learn how unsupervised techniques & shifts in the normative distribution affect Alzheimer's Disease detection🤯. Join me on Oct 12, 2:50 PDT in #Vancouver!🤝 @EstherPuyol Bene @DanielRueckert @ja_schnabel @AtoAndyKing Paper:

@cosmin_bercea @DanielRueckert You can find me and the poster in Ballroom C today! #ICML2023 #IMLH workshop

🌺 🌴 Are you attending #ICML2023 in Hawaii? Don't miss the opportunity to meet the amazing @ja_schnabel and @DanielRueckert at the IMLH workshop on July 28th and dive into our exciting research on #diffusion models for #AnomalyDetection 🧠💡 Code: https://t.co/srA0NyXr17

Just had an amazing seminar on anomaly detection and interpretable ML! It's great to give students a glimpse into the academic life. 💡#TUM #CompAi #seminar @ja_schnabel https://t.co/mJsSPuwMnr

Just wrapped up an incredible experience at #MIDL2023 in Nashville! 🎉 Huge thanks to all the amazing people I met and the insightful discussions we had. 🙌 If you missed my talk, you can catch it here: https://t.co/7VIbtxLfey #AI #MachineLearning #MedicalImaging

📢 Join me at 9:15 pm CEST at #MIDL2023 and discover how our new method RA ( https://t.co/S0rvnaSLCc) pushes the boundaries of Unsupervised Anomaly Detection.🧠🔬. Join me for insightful discussions, and if you're in Nashville, let's connect! 🤝 #AnomalyDetection #MedicalImaging

Introducing #ProfGPT, my new AI Avatar by @synthesiaIO! Getting closer towards fully automating myself :)

How can we tackle the noise paradox of diffusion models for anomaly detection? Make sure to keep an eye out for our upcoming work at ICML IMLH 2023! @DanielRueckert @ja_schnabel Project page: https://t.co/zlXefXZgSZ #ai #anomalydetection #diffusion

Greatly enjoying the guest lecture by ⁦@BernhardKainz1⁩ ⁦@UniFAU⁩ ⁦@BioMedIAICL⁩ as part of our Masterseminar series on Anomaly segmentation ⁦@TU_Muenchen⁩ ⁦⁦@aimlab_tum⁩ ⁦@HelmholtzIml⁩