19,997 machine learning datasets
19,997 dataset results
The Objects States Detection Dataset consists of images depicting everyday household objects in a number of different states. The ground-truth annotations involve the labels and bounding boxes spanning 18 object categories and 9 state classes. The object categories are: \textit{bottle, jar, tub, book, drawer, door, cup, mug, glass, bowl, basket, box, phone, charger, socket, towel, shirt} and \textit{newspaper}. The 9 state classes are: \textit{open, close, empty, containing something liquid (CL), containing something solid (CS), plugged, unplugged, folded} and \textit{unfolded}.
The FaQuAD dataset is a reading comprehension dataset designed for evaluating question-answering models. Let me provide you with details about different versions of FaQuAD:
The UAV Delievery dataset created to advance the research in drone delivery, contains trajectory details of UAV in different speed, altitude, and wind conditions. The dataset is created from the Truck based "Online Food Delivery Platform" using Open Air Traffic Simulator (ATS) with a UAVTrajectory.py plugin. A pre-processing step is used to select the deliveries under distance 5km, due to battery constraints of UAVs. After the pre-processing, the dataset consists of a total number of 6911 deliveries that are simulated and collected in log files.
The dataset consists of 1024 x 1024 bitmap (.bmp) images, each containing a 16 x 16 array of image patches. Each patch is sampled as 64x64 grayscale, with a canonical scale and orientation. For details of how the scale and orientation is established, please see the paper.
The penguin dataset is a collection of images of penguin colonies in Antarctica coming from the larger penguin watch project, which was setup with the purpose of monitoring their changes in population. The images are taken by fixed cameras in over 40 different locations, which have been capturing an image per hour for several years. In order to track the colony sizes, the number of penguins in each of the images in the dataset is required. So far, the penguin count has been done with the help of citizen scientists on the Penguin Watch site by Zooniverse, where interested users can place dots on top of the penguins. Here we release part of this data to the vision community in order to learn from the crowd-sourced dot-annotations to automatically annotate these images.
A large set of images of flowers
The study showed that the apple scab can be detected in the high-resolution RGB images in an early stage of its development. If two datasets, the early and advanced stages, are grouped together, the scab in the early stage is not visible after image resizing for CNN inputs 200-500px.
Sample data in the numpy array format (.npy) at the link https://zenodo.org/record/7189381#.Y0a2UHZBxD9. Satellite: Planetscope 3 meter
DPB-5L is a Multilingual KG dataset containing 5 KGs in English, French, Japanese, Greek, and Spanish. The dataset is used for the Knowledge Graph Completion and Entity Alignment task. DPB-5L (Japanese) is a subset of DPB-5L with Japanese KG.
Dataset to segmentize coughs
Projection of RibFrac CT dataset to a 2D plane to imitate X-Ray data for a total of 880 images with multi-label segmentation masks. The dataset contains fine-grained 92 individual labels of anatomical structures, which, when including super-classes, lead to a total of 166 labels in both lateral and frontal view.
Face detection and subsequent localization of facial landmarks are the primary steps in many face applications. Numerous algorithms and benchmark datasets have been introduced to develop robust models for the visible domain. However, varying conditions of illumination still pose challenging problems. In this regard, thermal cameras are employed to address this problem, because they operate on longer wavelengths. However, thermal face and facial landmark detection in the wild is an open research problem because most of the existing thermal datasets were collected in controlled environments. In addition, many of them were not annotated with face bounding boxes and facial landmarks. In this work, we present a thermal face dataset with manually labeled bounding boxes and facial landmarks to address these problems. The dataset contains 9,982 images of 147 subjects collected under controlled and uncontrolled conditions. As a baseline, we trained the YOLOv5 object detection model and its adap
Facial landmark detection is a cornerstone in many facial analysis tasks such as face recognition, drowsiness detection, and facial expression recognition. Numerous methodologies were introduced to achieve accurate and efficient facial landmark localization in visual images. However, there are only several works that address facial landmark detection in thermal images. The main challenge is the limited number of annotated datasets. In this work, we present a thermal face dataset with annotated face bounding boxes and facial landmarks. The dataset contains 2,556 thermal images of 142 individuals, where each thermal image is paired with the corresponding visual image. To the best of our knowledge, our dataset is the largest in terms of the number of individuals. In addition, our dataset can be employed for tasks such as thermal-to-visual image translation, thermal-visual face recognition, and others. We trained two models for the facial landmark detection task to show the efficacy of our
RTI International (RTI) generated 2,611 labeled point locations representing 19 different land cover types, clustered in 5 distinct agroecological zones within Rwanda. These land cover types were reduced to three crop types (Banana, Maize, and Legume), two additional non-crop land cover types (Forest and Structure), and a catch-all Other land cover type to provide training/evaluation data for a crop classification model. Each point is attributed with its latitude and longitude, the land cover type, and the degree of confidence the labeler had when classifying the point location. For each location there are also three corresponding image chips (4.5 m x 4.5 m in size) with the point id as part of the image name. Each image contains a P1, P2, or P3 designation in the name, indicating the time period. P1 corresponds to December 2018, P2 corresponds to January 2019, and P3 corresponds to February 2019. These data were used in the development of research documented in greater detail in “Deep
A dataset that consists of the demographics, triage category, symptoms, and comorbidities of COVID-19 patients.
STAR is a novel benchmark for Situated Reasoning, which provides 60K challenging situated questions in four types of tasks, 140K situated hypergraphs, symbolic situation descriptions and logic-grounded diagnosis for real-world video situations.
FHRMA is an open-source project for Fetal Heart Rate Morphological Analysis containing Matlab source code and datasets. As a sub-project, it includes a deep learning method and dataset for automatic identification of the maternal heart rate (MHR) and, more generally, false signals (FSs) on fetal heart rate (FHR) recordings. The challenge concerns particularly the FHR signal recorded with Doppler sensors, on which MHR interference and other FSs are particularly common, but the dataset also includes FHR recorded with scalp-ECG. The training and validation dataset contained 1030 expert-annotated periods (mean duration: 36 min) from 635 recordings. Labels consist of annotating each time sample as either 1: False signal; 0: True signal, or -1: do not know or irrelevant.
Data Approximatively 2 hours of videos were captured from 7 viewpoints during a professional basketball game.
CVGL Camera Calibration Dataset consists of 49 camera configurations with town 1 having 25 configurations while town 2 having 24 configurations. The parameters modified for generating the configurations include fov, x, y, z, pitch, yaw, and roll. Here, fov is the field of view, (x, y, z) is the translation while (pitch, yaw, and roll) is the rotation between the cameras. The total number of image pairs is 79, 320, out of which 18, 083 belong to Town 1 while 61, 237 belong to Town 2, the difference in the number of images is due to the length of the tracks.
MO7 dataset consists of 50,000 images with over 900 unique objects and over 18 classes. The dataset was collected in Missouri, Kansas, and Washington, 3 states known for their rural environment and challenging rural road conditions. The data consists of over 400 miles recorded in the 3 states combined . Our objective was to choose routes that are challenging in nature and underrepresented in other available datasets. Thus, we focused on unmarked roads, curvy roads, hills, and unpaved gravel roads, in addition to the availability of objects that are specific to rural areas such as agricultural machinery, small construction machinery used around farms, and farm animals.