This workshop serves as a platform for the MICCAI community to discuss and debate crucial aspects of safety and uncertainty in medical image applications.
Thank you to everyone who attended #UNSURE2024 in Marakkech #MICCAI2024. Both the oral and poster sessions were full of crowds.
Congratulations to @AnnaMWundram and @GutweinSimon for winning the best paper awards.
Looking forward to seeing everyone in Daejeon at #MICCAI2025.
Congratulations @GutweinSimon on receiving the Best Poster Presentation Award of the #MICCAI2024@UNSURE_Workshop 🎉 for the publication “Fishing in Uncertainty: Synthetic Contrastive Learning for Genetic Aberration Detection. Proceedings are online now➡️ https://t.co/JfC2Cnroat
Great keynote today at @UNSURE_Workshop and UQinMIA tutorial by @MichelDojat on multimodal aspects of Uncertainty Quantification. Thanks all that came early today to participate!
Really good crowd at our poster session!! Come by poster room and get to know more about recent developments in machine learning safety for medical image analysis!! #MICCAI2024
Then poster session in the conference session during coffee break.
After coffee we have an amazing keynote by: Prof. @MichelDojat
This followed by a long oral session.
We end the workshop with a panel discussion and closing ceremony.
Looking forward to seeing everyone.
#MICCAI2024 is still not over!!! Come by #UNSURE2024 workshop if you are interested in safety aspect of machine learning in medical imaging.
Room: Oliveraie in the conference center
We start with amazing tutorial sessions: https://t.co/QNaTl7LlZ0
📢 New publication to be presented at @UNSURE_Workshop at #MICCAI2024!
📃🦾 Improving the reliability of AI radiology report generation models through a quality control framework!
👀 Check out our research: https://t.co/nabGxXMXe0
🧵 (1/5)
#20: Information Bottleneck-based Feature Weighting for Enhanced Medical Image OoD Detection by Brayden Schott et al. #MICCAI2024
📜: https://t.co/qyebmNNnAP
📽️: https://t.co/LgyZq3uUoS
Proud to have our paper on typicality-based OOD detection presented at #MICCAI2024#UNSURE!💡 By measuring how typical an image is, instead of its likelihood - we’re pushing the boundaries of anomaly detection in medical imaging.
#19: INFORMER by Zixin Shu, @mreyesag, et al. #MICCAI2024
Tl,Dr: New QC method enhances interpretability and uncertainty estimation in multi-label classification on CheXpert dataset, achieving higher F1 scores.
📜: https://t.co/tMgKqW2z15
📽️: https://t.co/BrN6W0NUw3
📣 New publication to be presented at @UNSURE_Workshop#MICCAI2024!
⚠️Individual #outofdistribution detection methods have strengths and weaknesses.
💡 Combining complementary methods can mitigate their weaknesses!
🧐 Dive into our research: https://t.co/Hb0sAuSm9S
🧵(1/10)
#18: Conformal Prediction in cancer screening by Christopher Clark, @kalpathy1, et al. #MICCAI2024
Tl,Dr: It compares MC dropout and CP for uncertainty estimation in cervical cancer screening!
📜: https://t.co/NE8VeqPrL1
📽️: https://t.co/fTn3LPsr7S
#17: GLANCE by Xianze Ai et al., #MICCAI2024
Tl,Dr: The framework uses global and local noise correction to enhance multi-label classification in chest X-rays, improving diagnostic accuracy and reliability!
📜: https://t.co/9ZPWiGSBMt
📽️: https://t.co/HksyNol0c6
#15: Efficient Precision control by @vincent_blot28 et al. #MICCAI2024
Tl,Dr: It enhances precision-recall in image analysis for ovarian follicle quantification, improving F1-score without retraining!
📜: https://t.co/grPRlFYXfr
📽️: https://t.co/2OSwMAn53H
#14: Uncertainty Aware ViT by @FXErick2,
@BernhardKainz1 et al. #MICCAI2024
Tl,Dr: New stochastic ViT enhances uncertainty estimation and performance in medical imaging, improving reliability for critical applications!
📜: https://t.co/BqtlR1pzIS
📷: https://t.co/dZivnREHPB