Latent Pollution Model: The Hidden Carbon Footprint in 3D Image Synthesis

Contemporary developments in generative AI are rapidly transforming the field of medical AI. These developments have been predominantly driven by the availability of large datasets and high computing power, which have facilitated a significant increase in model capacity. Despite their considerable potential, these models demand substantially high power, leading to high carbon dioxide (CO2) emissions. Given the harm such models are causing to the environment, there has been little focus on the carbon footprints of such models. This study analyzes carbon emissions from 2D and 3D latent diffusion models (LDMs) during training and data generation phases, revealing a surprising finding: the synthesis of large images contributes most significantly to these emissions. We assess different scenarios including model sizes, image dimensions, distributed training, and data generation steps. Our findings reveal substantial carbon emissions from these models, with training 2D and 3D models comparable to driving a car for 10 km and 90 km, respectively. The process of data generation is even more significant, with CO2 emissions equivalent to driving 160 km for 2D models and driving for up to 3345 km for 3D synthesis. Additionally, we found that the location of the experiment can increase carbon emissions by up to 94 times, and even the time of year can influence emissions by up to 50%. These figures are alarming, considering they represent only a single training and data generation phase for each model. Our results emphasize the urgent need for developing environmentally sustainable strategies in generative AI.

Multi-Modal Dataset Creation for Federated~Learning with DICOM Structured Reports

Purpose: Federated training is often hindered by heterogeneous datasets due to divergent data storage options, inconsistent naming schemes, varied annotation procedures, and disparities in label quality. This is particularly evident in the emerging multi-modal learning paradigms, where dataset harmonization including a uniform data representation and filtering options are of paramount importance. Methods: DICOM structured reports enable the standardized linkage of arbitrary information beyond the imaging domain and can be used within Python deep learning pipelines with highdicom. Building on this, we developed an open platform for data integration and interactive filtering capabilities that simplifies the process of assembling multi-modal datasets. Results: In this study, we extend our prior work by showing its applicability to more and divergent data types, as well as streamlining datasets for federated training within an established consortium of eight university hospitals in Germany. We prove its concurrent filtering ability by creating harmonized multi-modal datasets across all locations for predicting the outcome after minimally invasive heart valve replacement. The data includes DICOM data (i.e. computed tomography images, electrocardiography scans) as well as annotations (i.e. calcification segmentations, pointsets and pacemaker dependency), and metadata (i.e. prosthesis and diagnoses). Conclusion: Structured reports bridge the traditional gap between imaging systems and information systems. Utilizing the inherent DICOM reference system arbitrary data types can be queried concurrently to create meaningful cohorts for clinical studies. The graphical interface as well as example structured report templates will be made publicly available.

Federated Foundation Model for Cardiac CT Imaging

Federated learning (FL) is a renowned technique for utilizing decentralized data while preserving privacy. However, real-world applications often involve inherent challenges such as partially labeled datasets, where not all clients possess expert annotations of all labels of interest, leaving large portions of unlabeled data unused. In this study, we conduct the largest federated cardiac CT imaging analysis to date, focusing on partially labeled datasets ($n=8,124$) of Transcatheter Aortic Valve Implantation (TAVI) patients over eight hospital clients. Transformer architectures, which are the major building blocks of current foundation models, have shown superior performance when trained on larger cohorts than traditional CNNs. However, when trained on small task-specific labeled sample sizes, it is currently not feasible to exploit their underlying attention mechanism for improved performance. Therefore, we developed a two-stage semi-supervised learning strategy that distills knowledge from several task-specific CNNs (landmark detection and segmentation of calcification) into a single transformer model by utilizing large amounts of unlabeled data typically residing unused in hospitals to mitigate these issues. This method not only improves the predictive accuracy and generalizability of transformer-based architectures but also facilitates the simultaneous learning of all partial labels within a single transformer model across the federation. Additionally, we show that our transformer-based model extracts more meaningful features for further downstream tasks than the UNet-based one by only training the last layer to also solve segmentation of coronary arteries. We make the code and weights of the final model openly available, which can serve as a foundation model for further research in cardiac CT imaging.

FUNAvg: Federated Uncertainty Weighted Averaging for Datasets with Diverse Labels

Federated learning is one popular paradigm to train a joint model in a distributed, privacy-preserving environment. But partial annotations pose an obstacle meaning that categories of labels are heterogeneous over clients. We propose to learn a joint backbone in a federated manner, while each site receives its own multi-label segmentation head. By using Bayesian techniques we observe that the different segmentation heads although only trained on the individual client's labels also learn information about the other labels not present at the respective site. This information is encoded in their predictive uncertainty. To obtain a final prediction we leverage this uncertainty and perform a weighted averaging of the ensemble of distributed segmentation heads, which allows us to segment "locally unknown" structures. With our method, which we refer to as FUNAvg, we are even on-par with the models trained and tested on the same dataset on average. The code is publicly available at https://github.com/Cardio-AI/FUNAvg.

Unconditional Latent Diffusion Models Memorize Patient Imaging Data

Generative latent diffusion models hold a wide range of applications in the medical imaging domain. A noteworthy application is privacy-preserved open-data sharing by proposing synthetic data as surrogates of real patient data. Despite the promise, these models are susceptible to patient data memorization, where models generate patient data copies instead of novel synthetic samples. This undermines the whole purpose of preserving patient data and may even result in patient re-identification. Considering the importance of the problem, surprisingly it has received relatively little attention in the medical imaging community. To this end, we assess memorization in latent diffusion models for medical image synthesis. We train 2D and 3D latent diffusion models on CT, MR, and X-ray datasets for synthetic data generation. Afterwards, we examine the amount of training data memorized utilizing self-supervised models and further investigate various factors that can possibly lead to memorization by training models in different settings. We observe a surprisingly large amount of data memorization among all datasets, with up to 41.7%, 19.6%, and 32.6% of the training data memorized in CT, MRI, and X-ray datasets respectively. Further analyses reveal that increasing training data size and using data augmentation reduce memorization, while over-training enhances it. Overall, our results suggest a call for memorization-informed evaluation of synthetic data prior to open-data sharing.

Investigating Data Memorization in 3D Latent Diffusion Models for Medical Image Synthesis

Generative latent diffusion models have been established as state-of-the-art in data generation. One promising application is generation of realistic synthetic medical imaging data for open data sharing without compromising patient privacy. Despite the promise, the capacity of such models to memorize sensitive patient training data and synthesize samples showing high resemblance to training data samples is relatively unexplored. Here, we assess the memorization capacity of 3D latent diffusion models on photon-counting coronary computed tomography angiography and knee magnetic resonance imaging datasets. To detect potential memorization of training samples, we utilize self-supervised models based on contrastive learning. Our results suggest that such latent diffusion models indeed memorize training data, and there is a dire need for devising strategies to mitigate memorization.

Surgical Phase and Instrument Recognition: How to identify appropriate Dataset Splits

Purpose: The development of machine learning models for surgical workflow and instrument recognition from temporal data represents a challenging task due to the complex nature of surgical workflows. In particular, the imbalanced distribution of data is one of the major challenges in the domain of surgical workflow recognition. In order to obtain meaningful results, careful partitioning of data into training, validation, and test sets, as well as the selection of suitable evaluation metrics are crucial. Methods: In this work, we present an openly available web-based application that enables interactive exploration of dataset partitions. The proposed visual framework facilitates the assessment of dataset splits for surgical workflow recognition, especially with regard to identifying sub-optimal dataset splits. Currently, it supports visualization of surgical phase and instrument annotations. Results: In order to validate the dedicated interactive visualizations, we use a dataset split of the Cholec80 dataset. This dataset split was specifically selected to reflect a case of strong data imbalance. Using our software, we were able to identify phases, phase transitions, and combinations of surgical instruments that were not represented in one of the sets. Conclusion: In order to obtain meaningful results in highly unbalanced class distributions, special care should be taken with respect to the selection of an appropriate split. Interactive data visualization represents a promising approach for the assessment of machine learning datasets. The source code is available at https://github.com/Cardio-AI/endovis-ml

Why is the winner the best?

International benchmarking competitions have become fundamental for the comparative performance assessment of image analysis methods. However, little attention has been given to investigating what can be learnt from these competitions. Do they really generate scientific progress? What are common and successful participation strategies? What makes a solution superior to a competing method? To address this gap in the literature, we performed a multi-center study with all 80 competitions that were conducted in the scope of IEEE ISBI 2021 and MICCAI 2021. Statistical analyses performed based on comprehensive descriptions of the submitted algorithms linked to their rank as well as the underlying participation strategies revealed common characteristics of winning solutions. These typically include the use of multi-task learning (63%) and/or multi-stage pipelines (61%), and a focus on augmentation (100%), image preprocessing (97%), data curation (79%), and postprocessing (66%). The "typical" lead of a winning team is a computer scientist with a doctoral degree, five years of experience in biomedical image analysis, and four years of experience in deep learning. Two core general development strategies stood out for highly-ranked teams: the reflection of the metrics in the method design and the focus on analyzing and handling failure cases. According to the organizers, 43% of the winning algorithms exceeded the state of the art but only 11% completely solved the respective domain problem. The insights of our study could help researchers (1) improve algorithm development strategies when approaching new problems, and (2) focus on open research questions revealed by this work.

Biomedical image analysis competitions: The state of current participation practice

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Medical Diffusion -- Denoising Diffusion Probabilistic Models for 3D Medical Image Generation

Recent advances in computer vision have shown promising results in image generation. Diffusion probabilistic models in particular have generated realistic images from textual input, as demonstrated by DALL-E 2, Imagen and Stable Diffusion. However, their use in medicine, where image data typically comprises three-dimensional volumes, has not been systematically evaluated. Synthetic images may play a crucial role in privacy preserving artificial intelligence and can also be used to augment small datasets. Here we show that diffusion probabilistic models can synthesize high quality medical imaging data, which we show for Magnetic Resonance Images (MRI) and Computed Tomography (CT) images. We provide quantitative measurements of their performance through a reader study with two medical experts who rated the quality of the synthesized images in three categories: Realistic image appearance, anatomical correctness and consistency between slices. Furthermore, we demonstrate that synthetic images can be used in a self-supervised pre-training and improve the performance of breast segmentation models when data is scarce (dice score 0.91 vs. 0.95 without vs. with synthetic data).