3,275 machine learning datasets
3,275 dataset results
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 ge
The TXL-PBC Dataset is a comprehensive collection of re-annotated and integrated cell images from multiple cell datasets. The main objective of this study is to perform sample reduction, re-labeling, and integration from BCCD and BCD datasets. Then, the original dataset is integrated with two new cell datasets, PBC dataset Peripheral Blood Cells and Raabin-WBC dataset Raabin White Blood Cells, to create a high-quality, sample balanced new dataset. We call it TXL-PBC dataset. We use the Labelimg tool Semi-automated labeling is performed using YOLOv8n.to annotate all the datasets. It is specifically designed for evaluating various object detection models, especially those that use the YOLO format.
The Peripheral Blood Cell} (PBC) dataset consists of 17,092 images. These images are further organized into the following eight groups: neutrophils, eosinophils, basophils, lymphocytes, monocytes, immature granulocytes (including promyelocytes, myelocytes, and metamyelocytes), erythroblasts, and platelets or thrombocytes. Each image is 360 x 363 pixels in size and is in JPG format, annotated by expert clinical pathologists. This dataset focuses on images of peripheral blood cells. For our newly introduced dataset, we have selected five types of white blood cells from this dataset.
We introduce a new synthetic test set named IS3 for interactive sound source localization. By leveraging diffusion models, we generate images containing multiple sounding objects. Any combination of sounding objects can appear in the same scene. Additionally, this dataset offers unusual scenes and unique combinations that are rarely found in nature, such as ‘a donkey playing a saxophone’ or ‘a sea lion on the snow’. This dataset provides both segmentation maps and bounding box information with class categories. IS3 includes 3240 images, resulting in 6480 unique audio-visual instances (with 2 objects per image) across 118 categories. This dataset can be used in below tasks: 1) Sound Source Localization 2) Audio-Visual Segmentation 3) Semantic Segmentation
Visual Haystacks (VHs) is a "visual-centric" Needle-In-A-Haystack (NIAH) benchmark specifically designed to evaluate the capabilities of Large Multimodal Models (LMMs) in visual retrieval and reasoning over sets of unrelated images. Unlike conventional NIAH challenges that center on text-related retrieval and understanding with limited anecdotal examples, VHs contains a much larger number of examples and focuses on "simple visual tasks", providing a more accurate reflection of LMMs' capabilities when dealing with extensive visual context.
The Foggy KITTI dataset extends the KITTI dataset to include challenging weather conditions, aiming to support research in real-world applications such as autonomous driving. It contains synthetic fog images with different levels of intensity and is divided into training and testing sets, providing a useful resource for developing and evaluating models in practical scenarios.
Click to add a brief description of the dataset (Markdown and LaTeX enabled).
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.
Noise of Web (NoW) is a challenging noisy correspondence learning (NCL) benchmark for robust image-text matching/retrieval models. It contains 100K image-text pairs consisting of website pages and multilingual website meta-descriptions (98,000 pairs for training, 1,000 for validation, and 1,000 for testing). NoW has two main characteristics: without human annotations and the noisy pairs are naturally captured. The source image data of NoW is obtained by taking screenshots when accessing web pages on mobile user interface (MUI) with 720 $\times$ 1280 resolution, and we parse the meta-description field in the HTML source code as the captions. In NCR (predecessor of NCL), each image in all datasets were preprocessed using Faster-RCNN detector provided by Bottom-up Attention Model to generate 36 region proposals, and each proposal was encoded as a 2048-dimensional feature. Thus, following NCR, we release our the features instead of raw images for fair comparison. However, we can not just
Click to add a brief description of the dataset (Markdown and LaTeX enabled).
Click to add a brief description of the dataset (Markdown and LaTeX enabled).
This work contributes a large, complex, and realistic high-quality safety clothing and helmet detection (SFCHD) dataset. The dataset comprises 12,373 images, covering 7 categories, with a total of 50,558 labeled instances. All images are captured from factory surveillance cameras, encompassing 40 different scenes across two chemical plants. It is worth noting that our SFCHD dataset not only provides a rich set of training samples but also serves as a benchmark for the evaluation of various detection tasks, such as small object detection, and high-low light object detection.
Early detection of plant nutritional deficiencies, followed by corrective actions, is essential for sustaining crop yield. However, identifying these early signs in plant leaves remains challenging, even with computer-aided diagnostic tools, due to their often subtle nature. To address this, we introduce a public dataset focused on three cucurbits: ash gourd, bitter gourd, and snake gourd. We chose to focus on these cucurbits due to their significant impact on global vegetable production. The dataset includes 2,700 segmented and augmented leaf images, capturing early indicators of nitrogen and potassium deficiencies alongside a healthy control group. This dataset, to the best of our knowledge, is the first to specifically target the critical early stage of nutritional stress in plants.
The success of any AI-driven system relies heavily on vast amounts of training data. While AI applications in plant stress management have gained attention in recent years, there's still a significant lack of expert-annotated data, especially for tropical and subtropical crops. We're filling in this gap by releasing a public dataset with 4,749 leaf pictures of healthy, nutrient-deficient, and pest-affected tomatoes, eggplants, cucumbers, bitter gourds, snake gourds, ridge gourds, ash gourds, and bottle gourds. This dataset encompasses 57 unique classes, with high-resolution images (3024 x 3024) captured at three different sites in Bangladesh under natural field conditions. An expert panel from the Bangladesh Agricultural Research Institute (BARI) has labeled the images. This collection not only features the largest number of plant stress classes but also introduces the first multi-label classification challenge in the agricultural domain.
The Modified Swiss Dwellings (MSD) dataset is an ML-ready dataset for floor plan generation and analysis at building-level scale. The MSD dataset is completely derived from the Swiss Dwellings database (v3.0.0). The MSD dataset contains highly-detailed 5372 floor plans of single- as well as multi-unit building complexes across Switzerland, hence extending the building scale w.r.t. of other well know floor plan datasets like the RPLAN dataset.
This dataset is a "part I" extension of the "Engineered cardiac microbundle time-lapse microscopy image dataset" and contains 732 experimental time-lapse image sequences of beating hiPSC-based cardiac microbundles using microbundle strain gauge platforms [1] ("Type1"). In "part II" extension, we include 808 experimental time-lapse image sequences of beating hiPSC-based cardiac microbundles using FibroTUG platforms [2] ("Type2").
This dataset is a "part II" extension of the "Engineered cardiac microbundle time-lapse microscopy image dataset" and contains 808 experimental time-lapse image sequences of beating hiPSC-based cardiac microbundles using FibroTUG platforms [1] ("Type2"). In "part I" extension, we include 732 experimental time-lapse image sequences of beating hiPSC-based cardiac microbundles using microbundle strain gauge platforms [2] ("Type1").
Mapping urban large-area advertising structures using drone imagery and deep learning-based spatial data analysis.
A configurable synthetic dataset of simple shapes with ground truth concepts and known causal relationships between concepts and classes.
An adaption of the MVTec Anomaly Detection dataset, presented in the paper "Domain-independent detection of known anomalies".