Datasets
Models
-
GeorgeSullivan
-
-
5 Follower(s)
18 Following
Datasets
Dataset for pneumonia classification using chest X-rays. Contains balanced train/test/pred folders for CNN model development. Pneumonia is lung inflammation from infection. Diagnostic methods include X-ray, CT, ultrasound, biopsy.
Image
Radiology
CT
Chest
Diagnosis
Classification
Other
1
7
BrixIA COVID-19 Project: BS-Net architecture for COVID-19 severity scoring on chest X-rays. Includes 4695 annotated DICOM CXRs with Brixia scores, segmentation datasets, and open-source code for research purposes.
1
publicly available multimodal dataset of 324 patient studies from Stanford University Hospital. Contains chest X-rays (JPG format) and clinical indications describing age, sex, and patient conditions.
Mixed
1
CT scan dataset for cardiomegaly detection containing DICOM images from 20 subjects: 11 with cardiomegaly and 9 healthy controls. Used for diagnostic classification of enlarged heart condition.
1
Chest cancer detection project using CNN to classify CT images into 3 cancer types (Adenocarcinoma, Large cell carcinoma, Squamous cell carcinoma) and normal. Dataset split: 70% train, 20% test, 10% validation. Images in JPG/PNG format.
1
Chest cancer detection project using CNN deep learning to classify and diagnose cancer types. Dataset includes Adenocarcinoma, Large cell carcinoma, Squamous cell carcinoma, and normal images in JPG/PNG format with 70/20/10 train/test/validation spli
1
Dataset of 500 patients with chest CT scans in .jpeg format for developing ML models to detect lung irregularities including nodules and lesions. Includes patient demographics and diagnostics from NIH Deep Lesion Dataset.
3
CT lung segmentation dataset with images and masks for lungs, heart, and trachea. Originally NRRD files converted to tensor format and RGB images. Contains training data with 3-class segmentation masks.
1
Chest X-ray dataset containing images in .jpg/.dcm formats organized by medical conditions (pneumonia, cardiomegaly, fractures, etc.) for ML research in automated detection and diagnosis of chest abnormalities.
1
COVID-19 chest X-ray dataset with 6,432 images in 3 categories (COVID19, PNEUMONIA, NORMAL). Organized in train/test folders (80/20 split). Collected from public sources for AI-based COVID-19 diagnosis applications.
1
Dataset for COVID-19 detection from chest X-rays containing COVID-positive and normal images in jpeg/jpg/png format. Compiled from existing datasets for learning purposes to detect COVID-19 from chest imaging.
1
Chest X-ray dataset with 5840 images for pneumonia detection using CNN. Contains train/test folders with NORMAL and PNEUMONIA subfolders for binary classification of chest X-ray images.
1
Dataset of 60 chest X-ray images with manual anatomical segmentation labels for lungs, heart, diaphragm, and spinal column. Created for Kaggle COVID-19 detection competition using BIMCV-COVID19 and MIDRC-RICORD data.
1
Chest X-ray dataset from Japan (~2000) with high-quality scanned films. Each case contains 1 nodule rated by 20 radiologists (AUC 0.72-0.89). Ideal for testing nodule detection performance across different subtlety levels.
1
Dataset of 5,863 chest X-ray images (JPEG) from pediatric patients (1-5 years) with 2 categories: Pneumonia/Normal. Organized in train/test/val folders. Quality-controlled and expert-graded for AI disease detection/classification.
1
Chest X-ray dataset with 7135 images in 4 categories (Normal/Pneumonia/Covid-19/Tuberculosis) organized in train/test/val folders. Sourced from multiple Kaggle datasets for AI-based disease detection and classification from X-ray images.
1
TB X-ray dataset with 500+ clinically labeled scans from China (Shenzhen: 662 images) & USA (Montgomery: 138 images). Enables automated tuberculosis diagnosis to assist radiologists in developing countries & improve screening efficiency.
1
Dataset of 100 chest X-ray studies: 50 with pulmonary nodules (6-30mm lung tumors), 50 normal. 100 patients (47M/51F), ages 18-87 (median 58). From Moscow research center, CC BY-NC-ND 3.0 licensed.
2
COVID-19 chest X-ray dataset for ML classification of healthy vs pneumonia patients. Includes SARS, Streptococcus & ARDS categories. Aims to develop AI application for faster virus detection using automated image analysis.
1
Large lung CT dataset for COVID-19 built from 7 public datasets. Contains 2729 image-mask pairs with COVID-19 lesion annotations from 3 datasets. All lesion types mapped to white for consistency across datasets.
1
This dataset contains the chest X-ray images of COVID-19 affected people and the Normal people.
1
COVID-19 chest X-ray database: 3616 COVID cases plus normal/pneumonia images. Qatar University & Dhaka researchers. Includes lung masks for AI screening research. PNG format, 299x299 pixels. Academic use.
1
It is an open database of COVID-19 cases with chest X-ray or CT images. It is designed for AI research to develop diagnostic and prognostic tools. It includes annotations.
1
COVID-19 chest X-ray segmentation dataset with 100 manually annotated images by radiologists. Includes anatomical structures (lungs, airways), pathologies (ground glass opacities, consolidation), and medical devices. COCO format.
1
COVID-19 chest imaging dataset with X-ray and CT images of COVID-19, MERS, SARS, and ARDS cases, including detailed metadata for AI diagnostic tool development.
1
CT scan chest dataset with 50,000+ studies with protocols and 100,000+ without, covering 24 pathologies including lung cancers and pneumonias. Designed for medical research to develop diagnostic tools and prediction models.
3
Single-cell RNA-sequencing study of lung tissue from healthy controls and systemic sclerosis-ILD patients. Analysis of 56,196 cells identified fibroblast heterogeneity and myofibroblast transcriptomes in disease progression.
Transcriptomics
1
Dataset of 5,856 chest X-ray images for pneumonia classification (NORMAL/BACTERIA/VIRUS) from pediatric patients aged 1-5 years. Images from Guangzhou Women and Children's Medical Center, split into training/testing sets for CNN model development.
1
Dataset derived from LUNA containing lung CT images categorized as healthy/unhealthy slices, with corresponding nodule masks and lung masks for training lung nodule detection and classification models.
1
LUNA16: 888 CT scans for lung nodule detection with 1186 annotated nodules >=3mm validated by 3+ radiologists. Includes MetaImage CT scans, annotations, candidates for false positive reduction, lung segmentation for 10-fold cross-validation.
1
COVID-19 chest X-ray dataset with 616 images containing lung bounding boxes. COCO format annotations with left/right lung boxes plus metadata (Finding, Modality, Sex, Survival, View). Manually annotated by radiologists from public sources.
1
IQ-OTH/NCCD lung cancer dataset: 1190 CT images from 110 cases collected in Iraq hospitals (fall 2019). Three classes: normal (55), benign (15), malignant (40). DICOM format, Siemens SOMATOM scanner, 1mm slices. De-identified data from middle Iraq pa
1
IQ-OTH/NCCD lung cancer dataset: 1190 CT scan images from 110 cases (55 normal, 15 benign, 40 malignant). Collected from Iraqi hospitals in 2019. DICOM format, Siemens scanner, de-identified data for lung cancer diagnosis research.
1
MRI dataset with 1018 lung images annotated by 4 radiologists for lung nodule identification. 80/20 train/test split. Training data augmented via rotation, flipping, resizing, translation. Images resized to 256×256
1
Lung X-Ray Image Dataset contains 3,475 images across 3 classes: Normal (1250), Lung Opacity (1125), and Viral Pneumonia (1100). Used for detecting and diagnosing lung diseases to improve patient outcomes and public health.
3
Dataset describing lung diseases with patient clinical data. Contains disease information (name, type, symptoms, causes, treatment), vital signs (heart rate, blood pressure, temperature), and infection status for diseases like tuberculosis, pneumonia
Diagnoses
3
Dataset for lung cancer node classification as malignant/benign. Contains medical images in TIFF and HDF5 formats from LUNA16 competition. Used for practicing CNN approaches vs manual feature extraction methods.
1
Lung Nodule Segmentation Dataset with high-resolution CT scans and expert annotations. Contains 239 training and 41 test images (416x416 PNG format) with corresponding segmentation masks for ML model development in lung cancer detection.
2
Dataset of 112,120 chest X-ray images with 14 disease labels from 30,805 patients, used for multi-class classification of thoracic pathologies like pneumonia, atelectasis, consolidation, and other lung conditions.
1
89 NSCLC patients with CT scans, gene expression, and clinical data. Radiomics study extracting 440 tumor features for radiogenomics analysis associating imaging features with gene expression patterns.
1
PadChest: Large dataset of 160K+ chest X-ray images from 67K patients with radiologist reports. Labeled with 174 findings, 19 diagnoses, 104 locations using manual annotation and RNN. First public Spanish X-ray database.
3
Pediatric chest X-ray dataset with 5,856 images labeled as pneumonia or normal. Addresses challenges of acquiring standardized pediatric X-rays and captures different physiology compared to adult-focused datasets.
1
Dataset: 807 thoracic CT images from 313 patients (2016-2022, Firat University). Images show pulmonary artery bifurcation. Patients grouped by pulmonary hypertension severity: Control (114pts, 210imgs), Group1 (43pts, 80imgs), Group2 (65pts, 130imgs)
1
RadGenome Chest CT: Large-scale 3D chest CT dataset with 25,692 volumes from 20,000 patients. Features 197-category organ segmentation, 665K grounded reports, 1.3M VQA pairs for multimodal medical AI development.
1
920 annotated audio recordings from 126 patients. Contains crackles, wheezes, and clean sounds recorded via digital stethoscopes. Includes demographic data. Used for ML-based diagnosis of asthma, COPD, pneumonia.
3
3
Author retrained models on chest X-ray dataset as better starting points than ImageNet. ResNet200D achieved 96.7% CV/LB. Need to add classifier layer when loading. Won't share weights to keep leaderboard stable. Data: NIH chest X-rays.
1
TB chest X-ray database: 700 public TB images, 2800 via NIAID portal, 3500 normal images. Created by Qatar University/University of Dhaka. Achieved 98.3% deep learning classification accuracy.
1
Preprocessed UniToChest dataset containing CT lung scans with Hounsfield Units conversion, windowing, U-Net lung segmentation, and CLAHE contrast enhancement. Includes large (>10mm) and small (<10mm) nodule subsets for cancer detection research.
1
COVID-19 chest X-ray dataset with 6500 images, 517 COVID cases. Includes pixel-level lung segmentations, pneumonia classification (viral/bacterial/fungal/healthy), and patient metadata (age, sex, outcome). Multi-source data with varying resolutions.
1
Chest X-ray dataset with 9 lung disease categories: Normal, Inflammatory, Higher/Lower density, Obstructive, Degenerative infectious, Encapsulated lesions, Mediastinal/Chest changes. Synthetically augmented, de-identified images.
3