Medical Image Segmentation

Click to add a brief description of the dataset (Markdown and LaTeX enabled). Provide: * a high-level explanation of the dataset characteristics * explain motivations and summary of its content * potential use cases of the dataset

Click to add a brief description of the dataset (Markdown and LaTeX enabled). Provide: * a high-level explanation of the dataset characteristics * explain motivations and summary of its content * potential use cases of the dataset

Several datasets are fostering innovation in higher-level functions for everyone, everywhere. By providing this repository, we hope to encourage the research community to focus on hard problems. In this repository, we present our medical imaging DICOM files of patients from our User Tests and Analysis 4 (UTA4) study. Here, we provide a dataset of the used medical images during the UTA4 tasks. This repository and respective dataset should be paired with the dataset-uta4-rates repository dataset. Work and results are published on a top Human-Computer Interaction (HCI) conference named AVI 2020 (page). Results were analyzed and interpreted on our Statistical Analysis charts. The user tests were made in clinical institutions, where clinicians diagnose several patients for a Single-Modality vs Multi-Modality comparison. For example, in these tests, we used both prototype-single-modality and prototype-multi-modality repositories for the comparison. On the same hand, the hereby dataset represents the pieces of information of both BreastScreening and MIDA projects. These projects are research projects that deal with the use of a recently proposed technique in literature: Deep Convolutional Neural Networks (CNNs). From a developed User Interface (UI) and framework, these deep networks will incorporate several datasets in different modes. For more information about the available datasets please follow the Datasets page on the Wiki of the meta information repository. Last but not least, you can find further information on the Wiki in this repository. We also have several demos to see in our YouTube Channel, please follow us.

BRISC is a high-quality, expert-annotated MRI dataset curated for brain tumor segmentation and classification. It addresses common limitations in existing datasets (e.g., BraTS, Figshare), including class imbalance, narrow tumor focus, and annotation inconsistencies.

Highlights • Publicly available dataset with 82 H&E stained images of frozen sections. • Images are acquired on 19 patients with metastatic colon cancer in a liver. • Originally stained and two stain-normalized sets of images included • Pixel wise ground truths provided by seven domain experts.

This is the dataset released along with the publication: CUTS: A Deep Learning and Topological Framework for Multigranular Unsupervised Medical Image Segmentation [ArXiv] [MICCAI] [GitHub]

HeLa cells on a flat glass Dr. G. van Cappellen. Erasmus Medical Center, Rotterdam, The Netherlands

ENSeg Dataset Overview This dataset represents an enhanced subset of the ENS dataset. The ENS dataset comprises image samples extracted from the enteric nervous system (ENS) of male adult Wistar rats (Rattus norvegicus, albius variety), specifically from the jejunum, the second segment of the small intestine. The original dataset consists of two classes: - Control (C): Healthy animals. - Walker-256 Tumor (WT): Animals with cancer induced by the Walker-256 Tumor. Image acquisition involved 13 different animals, with 7 belonging to the C class and 6 to the WT class. Each animal contributed 32 image samples obtained from the myenteric plexus. All images were captured using the same setup and configuration, stored in PNG format, with a spatial resolution of 1384 × 1036 pixels. The overall process of obtaining the images and performing the morphometric and quantitative analyses takes approximately 5 months. ### Dataset Annotations Our dataset version includes expert-annotated labels for 6 animals, tagged as 2C, 4C, 5C, 22WT, 23WT, and 28WT. The image labels were created by members of the same laboratory where the images originated: researchers from the Enteric Neural Plasticity Laboratory of the State University of Maringá (UEM). Annotations were generated using LabelMe and consist of polygons marking each neuron cell. To maintain labeling quality according to laboratory standards, only neuron cells with well-defined borders were included in the final label masks. The labeling process lasted 9 months (from November 2023 to July 2024) and was iteratively reviewed by the lead researcher of the lab. ### Dataset Statistics After processing, the full dataset contains: - 187 images - 9,709 annotated neuron cells The table below summarizes the number of images and annotated neurons per animal tag. | Animal Tag | # of Images | # of Neurons | |------------|------------|--------------| | 2C | 32 | 1590 | | 4C | 31 | 1513 | | 5C | 31 | 2211 | | 22WT | 31 | 1386 | | 23WT | 31 | 1520 | | 28WT | 31 | 1489 | | Total | 187 | 9709 | ### Recommended Training Methodology Due to the limited number of animal samples and the natural split of images, we recommend using a leave-one-out cross-validation (LOO-CV) method for training. This ensures more reliable results by performing six training sessions per experiment, where: - In each session, images from one subject are isolated for testing. - The remaining images are used for training. Although cross-validation is recommended, it is not mandatory. Future works may introduce new training methodologies as the number of annotated subjects increases. However, it is crucial to maintain images from the same source (animal) within the same data split to prevent biased results. Even though samples are randomly selected from the animals' original tissues, this approach enhances the credibility of the findings. This dataset provides a valuable resource for instance segmentation and biomedical image analysis, supporting research on ENS morphology and cancer effects. Contributions and feedback are welcome! ## Project Citation If you want to cite our article, the dataset, or the source codes contained in this repository, please used the citation (bibtex format): @Article{felipe25enseg, AUTHOR = { Felipe, Gustavo Zanoni and Nanni, Loris and Garcia, Isadora Goulart and Zanoni, Jacqueline Nelisis and Costa, Yandre Maldonado e Gomes da}, TITLE = {ENSeg: A Novel Dataset and Method for the Segmentation of Enteric Neuron Cells on Microscopy Images}, JOURNAL = {Applied Sciences}, VOLUME = {15}, YEAR = {2025}, NUMBER = {3}, ARTICLE-NUMBER = {1046}, URL = {https://www.mdpi.com/2076-3417/15/3/1046}, ISSN = {2076-3417}, DOI = {10.3390/app15031046} } ## Additional Notes Please check out our previous datasets if you are interest into developing projects with Enteric Nervous System images: 1. EGC-Z: three datasets of Enteric Glial cells images, composed of three different chronic degenerative diseases: Cancer, Diabetes Mellitus, and Rheumatoid Arthritis. Each dataset represent binary classification task, with the classes: control (healthy) and sick; 2. ENS: the ENS image datasets comprises 1248 images taken from thirteen rats distributed in two classes: control/healthy or sick. The images were created with three distinct contrast settings targeting different Enteric Nervous System cells: Enteric Neuron cells, Enteric Glial cells, or both. For more details, please contact the main author of this project or create an issue on the project.

A dataset of abdominal CT studies in NifTi format from the open-source medical data repository Medical Decathlon was utilized. To expedite the partitioning process, the MONAILabel plugin of the MONAI framework within the 3D Slicer program was employed. A radiologist with 15 years of experience conducted a validation process, wherein the boundaries of the colon markup were verified on each slice. The existing colorectal cancer markings in the dataset remained unaltered. Validation by a radiologist reduced the size of the validated dataset to 122 studies. In this case, the 122 studies were categorized into three subsets based on the quality of the data: The "good" subset comprises 100 studies, while the "bad" subset contains 17 cropped studies (in which the entire colon is not visible on the image). The "bad" subset comprises five studies. Two of these studies were of poor quality and could not identify the entire colon. Two further studies involved colon stomas following surgery, while one study involved a hernia. Four studies were excluded from the set due to broken images, with axial plane images duplicated or the slice order mixed

MDA231 human breast carcinoma cells infected with a pMSCV vector including the GFP sequence, embedded in a collagen matrix Dr. R. Kamm. Dept. of Biological Engineering, Massachusetts Institute of Technology, Cambridge MA (USA)

Simulated nuclei of HL60 cells stained with Hoescht Dr. V. Ulman and Dr. D. Svoboda. Centre for Biomedical Image Analysis (CBIA), Masaryk University, Brno, Czech Republic (Created using MitoGen, part of Cytopacq)

A dataset of A 3D Computed Tomography (CT) image dataset, ImageTBAD, for segmentation of Type-B Aortic Dissection is published. ImageTBAD contains 100 3D Computed Tomography (CT) images, which is of decent size compared with existing medical imaging datasets. ImageTBAD contains a total of 100 3D CTA images gathered from Guangdong Peoples' Hospital Data from January 1,2013 to April 23, 2019. Images are acquired from a variety of scanners (GE Medical Systems, Siemens, Philips), resulting in large variance in voxel size, resolution and imaging quality. All the images are pre-operative TBAD CTA images whose top and bottom are around the neck and the brachiocephalic vessels, respectively, in the axial view. The segmentation labeling is performed by a team of two cardiovascular radiologists who have extensive experience with TBAD. The segmentation labeling of each patient is fulfilled by one radiologist and checked by the other. The segmentation

This dataset contains 65 DFIs acquired from patients with POAG at the University of Iowa Hospitals and Clinics. DFIs were acquired using a 30° Zeiss fundus camera (Niemeijer et al 2011). The images were centered on the optic disc. The original DFIs resolution was 2392 × 2048. In order to benchmark LUNet on this dataset, the black border of the DFIs were padded to a squared resolution of 2048 × 2048 pixels and then resized to a 1444 × 1444 pixels resolution. From the resulting DFIs, 15 optic disc-centered DFIs were randomly selected to form the second external test set. No other additional metadata were provided in the open source dataset.

This data set comprises 22 fundus images with their corresponding manual annotations for the blood vessels, separated as arteries and veins. It also include labels for glaucomatous / healthy, differentiating between normal tension glaucoma (NAG) and primary open angle glaucoma (POAG). Further information about the data set is provided in our MICCAI 2018 paper: Orlando, J. I., Breda, J. B., van Keer, K., Blaschko, M. B., Blanco, P. J., & Bulant, C. A. (2018). Towards a glaucoma risk index based on simulated hemodynamics from fundus images. MICCAI 2018. Please, cite the corresponding paper in case you use this data on a scientific publication. Any commercial usage is forbidden. This data set can only be used for scientific purposes.

Pulmonary hypertension (PH) is a syndrome complex that accompanies a number of diseases of different etiologies, associated with basic mechanisms of structural and functional changes of the pulmonary circulation vessels and revealed pressure increasing in the pulmonary artery. The structural changes in the pulmonary circulation vessels are the main limiting factor determining the prognosis of patients with PH. Thickening and irreversible deposition of collagen in the pulmonary artery branches walls leads to rapid disease progression and a therapy effectiveness decreasing. In this regard, histological examination of the pulmonary circulation vessels is critical both in preclinical studies and clinical practice. However, measurements of quantitative parameters such as the average vessel outer diameter, the vessel walls area, and the hypertrophy index claimed significant time investment and the requirement for specialist training to analyze micrographs. A dataset of pulmonary circulation vessels for pathology assessment using semantic segmentation techniques based on deep-learning is presented in this work. 609 original microphotographs of vessels, numerical data from experts’ measurements, and microphotographs with outlines of these measurements for each of the vessels are presented. The presented database will be useful for the development of new software solutions for the analysis of histological micrograph.

Different types of cells play a vital role in the initiation, development, invasion, metastasis and therapeutic response of tumors of various organs. For example, (1) most carcinomas originate from epithelial cells, (2) spatial arrangement of tumor infiltrating Lymphocytes (TILs) is associated with clinical outcome in several cancers, including the ones of breast, prostate, and lung (Fridman et. al., Nature Reviews Cancer, 2012), and (3) tumor associated macrophages (TAMs) influence diverse processes such as angiogenesis, neoplastic cell mitogenesis, antigen presentation, matrix degradation, and cytotoxicity in various tumors (Ruffel and Coussens, Cancer Cell, 2015). Thus, accurate identification and segmentation of nuclei of multiple cell-types is important for AI enabled characterization of tumor and its microenvironment. In this challenge, participants will be provided with H&E stained tissue images of four organs with annotations of multiple cell-types including epithelial cells, lymphocytes, macrophages, and neutrophils. Participants will use the annotated dataset to develop computer vision algorithms to recognize these cell-types from the tissue images of unseen patients released in the testing set of the challenge. Additionally, all cell-types will not have equal number of annotated instances in the training dataset which will encourage participants to develop algorithms for learning from imbalanced classes in a few shot learning paradigm. H&E staining of human tissue sections is a routine and most common protocol used by pathologists to enhance the contrast of tissue sections for tumor assessment (grading, staging, etc.) at multiple microscopic resolutions. Hence, we will provide the annotated dataset of H&E stained digitized tissue images of several patients acquired at multiple hospitals using one of the most common 40x scanner magnification. The annotations will be done with the help of expert pathologists.

1、 Competition name: The 2nd China Society of Image and Graphics (CSIG) Image and Graphics Technology Challenge: MRSpineSeg Challenge: Automated Multi-class Segmentation of Spinal Structures on Volumetric MR Images. 2、 Purpose： Degenerative spine diseases (e.g., lumbar disc herniation, spinal stenosis, etc.) have become important diseases affecting the health and quality of life of the elderly and working people. These degenerative spinal diseases often cause changes in the structural morphology and mechanical systems of the spine, resulting in pain, such as lumbar disc herniation, reduced disc height, and nerve compression. Magnetic resonance imaging (MRI) as a non-invasive examination method, it has good soft tissue imaging and no radiation. It is a reliable screening method for degenerative spine diseases. In clinical practice, the treatment of degenerative spinal disorders depends largely on physicians’ experience and lacks accurate quantitative analysis tools. 3D automatic segmentation (MR) images of multiclass spinal structures by MRI are a prerequisite for 3D reconstruction of spinal structures. It can provide quantitative analysis tools for building biomechanical models of the spine, simulating stresses in spinal structures, and assessing the prognosis of different treatment options for degenerative spinal diseases. This competition aims to gather global developers to explore efficient and accurate 3D automatic segmentation of spinal structure in MR images by using artificial intelligence technology. The spinal structure to be segmented includes 10 vertebrae and 9 intervertebral discs. 3、 Organizer: Qianjin,Feng, School of Biomedical Engineering, Southern Medical University, Guangdong Key Laboratory of medical image processing, China. 4、 Requirements for competition participants： It is open to the whole society. Personnel from colleges and universities, scientific research institutions and enterprises can sign up for the competition. The maximum number of each team is four. Each person can only participate in one team. After team registration, the team information cannot be changed. Note: all personnel who have access to the competition data are prohibited from participating in the competition. Those who have not access to the competition data of Southern Medical University can also participate in the competition. Southern Medical University has the right of final interpretation. 5、 Timeline: Registration (March 30, 2021 – May 22, 2021) 6、Citation: [1] Shumao Pang, Chunlan Pang, Lei Zhao, Yangfan Chen, Zhihai Su, Yujia Zhou, Meiyan Huang, Wei Yang, Hai Lu, Qianjin Feng*. SpineParseNet: Spine Parsing for Volumetric MR Image by a Two-Stage Segmentation Framework with Semantic Image Representation [J]. IEEE Transactions on Medical Imaging, 2021, 40(1): 262-273. [2] Shumao Pang, Chunlan Pang, Zhihai Su, Liyan Lin, Lei Zhao, Yangfan Chen, Yujia Zhou, Hai Lu, Qianjin Feng*. DGMSNet: Spine Segmentation for MR Image by a Detection-Guided Mixed-supervised Segmentation Network [J]. Medical Image Analysis, 2022, 102261.

DOI Mouse Brain MRI atlas (both in-vivo and ex-vivo) (repository relocated from the original webpage) ## List of atlases - FVB_NCrl: Brain MRI atlas of the wild-type FVB_NCrl mouse strain (used as the background strain for the rTg4510 which is a tauopathy model mice express a repressible form of human tau containing the P301L mutation that has been linked with familial frontotemporal dementia.) - NeAt: Brain MRI atlas of the whld-type C57BL/6J mouse strain. Atlas was created based on the original MRM NeAt mouse brain atlas (template images reoriented and bias-corrected, left/right structure label seperated, and 4th ventricle manual segmentation added). - Tc1 Cerebellum: TC1 mouse cerebellar cortical sublayer lobules.This mouse cerebellar atlas can be used for mouse cerebellar morphometry. ## Citation - If you use the segmented brain structure, or use the atlas along with the automatic mouse brain MRI segmentation tools, we ask you to kindly cite the following papers: - Ma D, Cardoso MJ, Modat M, Powell N, Wells J, Holmes H, Wiseman F, Tybulewicz V, Fisher E, Lythgoe MF, Ourselin S. Automatic structural parcellation of mouse brain MRI using multi-atlas label fusion. PloS one. 2014 Jan 27;9(1):e86576. http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0086576 - Ma D, Holmes HE, Cardoso MJ, Modat M, Harrison IF, Powell NM, O'Callaghan J, Ismail O, Johnson RA, O’Neill MJ, Collins EC, Mirza F. Beg, Karteek Popuri, Mark F. Lythgoe, and Sebastien Ourselin Study the longitudinal in vivo and cross-sectional ex vivo brain volume difference for disease progression and treatment effect on mouse model of tauopathy using automated MRI structural parcellation. Frontiers in Neuroscience. 2019;13:11. https://www.frontiersin.org/articles/10.3389/fnins.2019.00011 - If you use the brain MR images of the FVB_NCrl mouse strain (the wildtype background of rTg4510), we ask you to kindly cite the following papers: - Wells JA, O'Callaghan JM, Holmes HE, Powell NM, Johnson RA, Siow B, Torrealdea F, Ismail O, Walker-Samuel S, Golay X, Rega M. In vivo imaging of tau pathology using multi-parametric quantitative MRI. Neuroimage. 2015 May 1;111:369-78. https://www.sciencedirect.com/science/article/pii/S105381191500124X - Holmes HE, Colgan N, Ismail O, Ma D, Powell NM, O'Callaghan JM, Harrison IF, Johnson RA, Murray TK, Ahmed Z, Heggenes M. Imaging the accumulation and suppression of tau pathology using multiparametric MRI. Neurobiology of aging. 2016 Mar 1;39:184-94. https://www.sciencedirect.com/science/article/pii/S0197458015006053 - Holmes HE, Powell NM, Ma D, Ismail O, Harrison IF, Wells JA, Colgan N, O'Callaghan JM, Johnson RA, Murray TK, Ahmed Z. Comparison of in vivo and ex vivo MRI for the detection of structural abnormalities in a mouse model of tauopathy. Frontiers in neuroinformatics. 2017 Mar 31;11:20. https://www.frontiersin.org/articles/10.3389/fninf.2017.00020/full - If you're using the mouse MRI T2* Active Starining Cerebellar atlas, we ask you to please kindly cite the following papers: - Ma, D., Cardoso, M. J., Zuluaga, M. A., Modat, M., Powell, N. M., Wiseman, F. K., Cleary, J. O., Sinclair, B., Harrison, I. F., Siow, B., Popuri, K., Lee, S., Matsubara, J. A., Sarunic, M. V, Beg, M. F., Tybulewicz, V. L. J., Fisher, E. M. C., Lythgoe, M. F., & Ourselin, S. (2020). Substantially thinner internal granular layer and reduced molecular layer surface in the cerebellum of the Tc1 mouse model of Down Syndrome – a comprehensive morphometric analysis with active staining contrast-enhanced MRI. NeuroImage, 117271. https://doi.org/https://doi.org/10.1016/j.neuroimage.2020.117271 - Ma, D., Cardoso, M. J., Zuluaga, M. A., Modat, M., Powell, N., Wiseman, F., Tybulewicz, V., Fisher, E., Lythgoe, M. F., & Ourselin, S. (2015). Grey Matter Sublayer Thickness Estimation in the Mouse Cerebellum. In Medical Image Computing and Computer Assisted Intervention 2015 (pp. 644–651). https://doi.org/10.1007/978-3-319-24574-4_77 ## Reference - For the original information of the NeAt atlas, please please refer to the website: http://brainatlas.mbi.ufl.edu/, and the following two reference papers: - Ma Yu, Smith David, Hof Patrick R, Foerster Bernd, Hamilton Scott, Blackband Stephen J, Yu Mei, Benveniste Helene In Vivo 3D Digital Atlas Database of the Adult C57BL/6J Mouse Brain by Magnetic Resonance Microscopy. Front. Neuroanat. 2, 1 (2008). - Ma Yu, Hof P R, Grant S C, Blackband S J, Bennett R, Slatest L, McGuigan M D, Benveniste H A three-dimensional digital atlas database of the adult C57BL/6J mouse brain by magnetic resonance microscopy. Neuroscience 135, 1203–15 (2005). ## Funding The works in this repositories received multiple funding from EPSRC, UCL Leonard Wolfson Experimental Neurology center, Medical Research Council (MRC), the NIHR Biomedical Research Unit (Dementia) at UCL and the National Institute for Health Research University College London Hospitals Biomedical Research center, the UK Regenerative Medicine Platform Safety Hub, and the Kings College London and UCL Comprehensive Cancer Imaging center CRUK & EPSRC in association with the MRC and DoH (England), UCL Faculty of Engineering funding scheme, Alzheimer Society Reseasrch Program from Alzheimer Society Canada, NSERC, CIHR, MSFHR Canada， Eli Lilly and Company, Wellcome Trust, the Francis Crick Institute, Cancer Research UK, and University of Melbourne McKenzie Fellowship.

This dataset was built with data acquired at the Hospital Clinic of Barcelona, Spain. It is composed of a total of 1126 HD polyp images. There are a total of 473 unique polyps, with a variable number of different shots per polyp (minimum: 2, maximum: 24, median: 10). Special attention was paid to ensure that images from the same polyp show different conditions. An external frame-grabber and a white light endoscope were used to capture raw images. The dataset contains images with two different resolutions: 1920 x 1080 and 1350 x 1080.

A total of 227 cross sectional images (20 x 54 mm with a resolution of 289 x 648 pixels) of hind-leg xenograft tumors from 29 mice were obtained with 1mm step-wise movement of the array mounted on a manual positioning device. The whole tumor volume was acquired using a diagnostic ultrasound system with a 10 MHz linear transducer and 50 MHz sampling. The dataset images are in Analyze 7.5 file format consisting of two files: a header with information about the size and number of voxels in each dimension (filename extension .hdr), and the actual data in binary format (filename extension .img). The images in the dataset are formatted according to the following protocol: CASE NUMBER | FILE TYPE (LOC OR SEG) | FILE EXTENSION | where loc is the original image and seg is the associated segmentation mask. The volumetric data are categorised into progressive versus non-progressive cases in response to chemotherapy treatment. The example_case folder contains an output of the 3D Multi-fractal Nakagami Feature descriptor (MNF) algorithm when applied to one of the liver tumour cases, see below. The group of slice images for a particular case where transformed to corresponding Nakagami shape and scale parametric images, and animated GIF images representing the fractal slice maps and a video of the corresponding fractal volume maps were estimated and included as well.

Rethinking Abdominal Organ Segmentation (RAOS) in the clinical scenario: A robustness evaluation benchmark with challenging cases. This dataset consists of 413 real clinical CT scans and 413x9 MR scans, all 19 organs were annotated by a senior oncologist (MD. Wenjun Liao, 10 years experiment). some challenging cases in clinical practice can help evaluate the generalization and robustness of deep learning methods. Some organ annotations not present in previous public datasets (such as prostate, seminal vesicles, etc).

A dataset made of 3D image data and their embeddings to test TomoSAM

We introduce the trapped yeast cell (TYC) dataset, a novel dataset for understanding instance-level semantics and motions of cells in microstructures. We release $105$ dense annotated high-resolution brightfield microscopy images, including about $19$k instance masks. We also release $261$ curated video clips composed of $1293$ high-resolution microscopy images to facilitate unsupervised understanding of cell motions and morphology.

Microscopy is a cornerstone of biomedical research, enabling detailed study of biological structures at multiple scales. Advances in cryo-electron microscopy, high-throughput fluorescence microscopy, and whole-slide imaging allow the rapid generation of terabytes of image data, which are essential for fields such as cell biology, biomedical research, and pathology. These data span multiple scales, allowing researchers to examine atomic/molecular, subcellular/cellular, and cell/tissue-level structures with high precision. A crucial first step in microscopy analysis is interpreting and reasoning about the significance of image findings. This requires domain expertise and comprehensive knowledge of biology, normal/abnormal states, and the capabilities and limitations of microscopy techniques. Vision-language models (VLMs) offer a promising solution for large-scale biological image analysis, enhancing researchers’ efficiency, identifying new image biomarkers, and accelerating hypothesis generation and scientific discovery. However, there is a lack of standardized, diverse, and large-scale vision-language benchmarks to evaluate VLMs’ perception and cognition capabilities in biological image understanding. To address this gap, we introduce μ-Bench, an expert-curated benchmark encompassing 22 biomedical tasks across various scientific disciplines (biology, pathology), microscopy modalities (electron, fluorescence, light), scales (subcellular, cellular, tissue), and organisms in both normal and abnormal states. We evaluate state-of-the-art biomedical, pathology, and general VLMs on μ-Bench and find that: i) current models struggle in all categories, even for basic tasks such as distinguishing microscopy modalities; ii) current specialist models fine-tuned on biomedical data often perform worse than generalist models; iii) fine-tuning in specific microscopy domains can cause catastrophic forgetting, eroding prior biomedical knowledge encoded in their base model. iv) weight interpolation between fine-tuned and pre-trained models offers one solution to forgetting and improves general performance across biomedical tasks. We release μ-Bench under a permissive license to accelerate the research and development of microscopy foundation models.

An instance segmentation dataset of yeast cells in microstructures. The dataset includes 493 densely annotated microscopy images. For more information see the paper "An Instance Segmentation Dataset of Yeast Cells in Microstructures".

ALFI (Annotations for Label-Free Images) is a dataset of images and annotations for label-free microscopy imaging. It consists of 29 time-lapse image sequences with various annotations (pixel-wise segmentation masks, object-wise bounding boxes, and tracking information), made publicly available to the scientific community through figshare. For a detailed description, see "L. Antonelli, F. Polverino, A. Albu, A. Hada, I.A. Asteriti, F. Degrassi, G. Guarguaglini, L. Maddalena, M.R. Guarracino, ALFI: Cell cycle phenotype annotations of label-free time-lapse imaging data from cultured human cells, Scientific Data 10(677), 2023". https://www.nature.com/articles/s41597-023-02540-1

Simulated GFP-actin-stained A549 Lung Cancer cells embedded in a Matrigel matrix Dr. M. Maška and Dr. D. V. Sorokin. Centre for Biomedical Image Analysis (CBIA), Masaryk University, Brno, Czech Republic (Created using FiloGen, part of Cytopacq)

Developing Tribolium Castaneum embryo (3D cartographic projection) Dr. A. Jain. Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany

Standardized Multi-Channel Dataset for Glaucoma (SMDG-19) is a collection and standardization of 19 public datasets, comprised of full-fundus glaucoma images, associated image metadata like, optic disc segmentation, optic cup segmentation, blood vessel segmentation, and any provided per-instance text metadata like sex and age. This dataset is the largest public repository of fundus images with glaucoma.

The UBIRIS.v2 iris dataset contains 11,102 iris images from 261 subjects with 10 images each subject. The images were captured under unconstrained conditions (at-a-distance, on-the-move and on the visible wavelength), with realistic noise factors. Source: Constrained Design of Deep Iris Networks Image Source: https://arxiv.org/pdf/1905.09481.pdf