Papers with Code 아카이브

2D Semantic Segmentation

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
17
평가표
11
보존된 평가 기록
42

데이터셋

출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.

Indian Food Image Dataset

This dataset is an extremely challenging set of over 5000+ original India food images captured and crowdsourced from over 800+ urban and rural areas, where each image is manually reviewed and verified by computer vision professionals at ****DC Labs. ### Dataset Features - Dataset size : 5000+ - Captured by : Over 800+ crowdsource contributors - Resolution : 99% images HD and above (1920x1080 and above) - Location : Captured with 800+ cities accross India - Diversity : Various lighting conditions like day, night, varied distances, view points, dimlight etc. - Device used : Captured using mobile phones in 2020-2021 - Usage : Indian food classification, Dish classification, Food plate detection, etc. ### Available Annotation formats COCO, YOLO, PASCAL-VOC, Tf-Record To download full datasets or to submit a request for your dataset needs, please ping us at sales@datacluster.ai Visit www.datacluster.ai to know more. Note: All the images are manually captured and verified by a large contributor base on DataCluster platform

InfiniteRep

InfiniteRep is a synthetic, open-source dataset for fitness and physical therapy (PT) applications. It includes 1k videos of diverse avatars performing multiple repetitions of common exercises. It includes significant variation in the environment, lighting conditions, avatar demographics, and movement trajectories. From cadence to kinematic trajectory, each rep is done slightly differently -- just like real humans. InfiniteRep videos are accompanied by a rich set of pixel-perfect labels and annotations, including frame-specific repetition counts. The dataset features: + 100 videos per exercise, spanning 5 to 10 repetitions each (1,000 videos total) + 7 unique indoor scenes + Realistic environmental occlusion (+ corresponding labels) + Diverse lighting conditions + Varied body shape, skin tones, and clothing + Rich annotations for 2D and 3D supervision ## Exercises The dataset currently includes the following exercises: + Pushups + Alternating Bicep Curls (with dumbbells) + Delt Flys (with dumbbells) + Squats + Bird Dogs + Supermans + Bicycle Crunches + Leg Raises + Front Raises (with dumbbells) + Overhead Press (with dumbbells) ## Annotations The dataset includes the following annotations: + Bounding boxes + Segmentation masks + Keypoints + Joint angles (quaternions) + Percent occlusion + Avatar characteristics + Camera position + and more Want depth labels? They are not included in the dataset but we can send them to you. Email us at info@toinfinity.ai. ## Download Download the dataset: toinfinity.ai/infiniterep Github repo with additional documentation: https://github.com/toinfinityai/InfiniteRep ## Need more data? Infinity AI specializes in generating custom synthetic data. If you need more (or different data), drop us a line at info@toinfinity.ai (we read every email).

KITTI-360-SR

Scene Recognition is a problem, where a set of visible objects must be correctly associated with objects marked on a semantic map - this problem is also sometimes called a Data Association. Please note, that Scene Recognition in terms where the observed scene must be labeled in terms such as 'kitchen', 'bedroom' and so on is a different problem. Scene Recognition is used in different mobile robot localization problems, such as semantic mapping, global localization, kidnapping, loop-closing and so on. But problem is that there is no special dataset for that task (at least I found none). Such dataset should have the following features: - a Semantic map with globally marked objects; - scenes data (views from robot camera), where objects are also labeled and their IDs are correctly matched with IDs on the Semantic map. For full info please follow https://github.com/MoscowskyAnton/scene_recognition_kitti_360

Linear Equation Image Dataset

Description: <a href="https://gts.ai/dataset-download/linear-equation-image-dataset/" target="_blank">👉 Download the dataset here</a> The Linear Equation Image Dataset is designed to help solve high school-level algebraic problems using machine learning (ML). It provides extensive visual data, ideal for training models in equation recognition and solving. Download Dataset What’s New Expanded Image Dataset: The dataset now contains over 30,000 images, covering a wide array of linear equations with varying complexities. The generation of equations follows multiple randomization techniques, ensuring diversity in the visual representation. Data Diversity: Equations include both simple and complex forms, with some involving fractional coefficients, inequalities, or multi-variable formats to increase the challenge. The images also come in different resolutions, fonts, and formats (handwritten and digitally rendered) to further test ML algorithms’ robustness. Possible Use Cases Symbolic Equation Recognition: Train models to visually recognize equations and convert them into symbolic form. Equation Solving: Create ML models capable of solving linear equations through image recognition. Handwritten Recognition: Use this dataset for handwriting recognition, helping machines interpret handwritten linear equations. Educational Tools: Develop AI tutors or mobile apps that assist students in solving linear equations by merely taking a photo of the problem. Algorithm Training: Useful for those researching symbolic computation, this dataset allows for testing and improving various image-to-text and equation-solving algorithms. Enhanced Research Opportunities This dataset can be particularly useful for educational institutions, research teams, and AI developers focusing on enhancing problem-solving capabilities via machine learning and symbolic computation models. This dataset is sourced from Kaggle.

maadaa-FaEco Dataset

The dataset is organized into 24 typical scenarios, showcasing the richness of real-world environments, conditions, and objects. It is carefully curated to reflect diverse and realistic situations, allowing models to be tested and refined under a wide range of conditions. This dataset contains 33 meticulously fine-annotated sub-datasets. Each is labeled and annotated with a high degree of precision. These datasets offer a vast range of potential use cases, from object detection and segmentation to pose estimation and beyond. This dataset provides a unique context in various applications including personalized recommendations, virtual fittings, beauty AI, and product recognition. This extensive, versatile dataset is a significant resource that we believe will inspire a myriad of innovative solutions in the field. Source: https://maadaa.ai/datasets/open-dataset/fashion-ecommerce/

MapAI Dataset

MapAI: Precision in Building Segmentation Dataset The dataset comprises 7500 training images and 1500 validation images from Denmark. The test dataset is split into two tasks, where the first task (1368 images) is to segment the buildings only using aerial images. In contrast, the second task (978 images) allows using aerial images and lidar data. All data samples have a resolution of 500x500. The aerial images are RGB images, while the lidar data are rasterized. The ground truth masks have two classes, building, and background. Source: MapAI: Precision in Building Segmentaiton Image Source: Kartai

Multi-Spectral Leaf Segmentation

This dataset were acquired with the Airphen (Hyphen, Avignon, France) six-band multi-spectral camera configured using the 450/570/675/710/730/850 nm bands with a 10 nm FWHM. And acquired on the site of INRAe in Montoldre (Allier, France, at 46°20'30.3"N 3°26'03.6"E) within the framework of the “RoSE challenge” founded by the French National Research Agency (ANR). Images contains bean, with various natural weeds (yarrows, amaranth, geranium, plantago, etc) and sowed ones (mustards, goosefoots, mayweed and ryegrass) with very distinct characteristics in terms of illumination (shadow, morning, evening, full sun, cloudy, rain, ...) The ground truth is defined for each images with polygons around leafs boundaries: In addition, each polygons are labeled into crop or weed. (2020-06-11)

Primitive Shape Abstraction

Dataset: RGB-D Images for Real-World and Synthetic Object Scenes This dataset consists of both real-world and synthetic RGB-D images, designed for object detection, classification, and segmentation tasks, particularly for primitive shape recognition. ## Real-World Data - Objects: 50 distinct objects captured using a Kinect camera. - Scenes: The dataset includes both single objects and piles of objects stacked over each other. - Images: Approximately 300 RGB-D images have been collected. - Data Format: RGB images paired with corresponding depth images. ## Synthetic Data - Simulator: The synthetic dataset was automatically generated using the CoppeliaSim simulator. - Objects: Only primitive shapes, including cuboid, semisphere, sphere, cylinder, stick, ring, and cone. - Scenes: Objects are dropped over each other to form complex scenes. - Images: The dataset contains 10,000 unique RGB-D images generated by the simulator. ## Labels Both the real-world and synthetic datasets are labeled with geometric primitive shape classes and boundaries, including: - Cuboid - Semisphere - Sphere - Cylinder - Stick - Ring - Cone This dataset is designed to support research in grasp detection, object recognition, and scene understanding using RGB-D data.

SignboardText

This dataset contains images and annotations for scene text detection and recognition. It is made up of two parts: (1) 1,175 images manually labeled with a total of 59,588 text instances at the line and word levels; and (2) 929 signboard images collected from the VinText, Total-Text, and ICDAR15 datasets. Each text instance in the first part of our dataset has a quadrilateral bounding box and a ground truth character sequence associated with it. In the second part, images are selected if they contain signboards. This portion of the dataset comprises 20,261 text instances at word levels. This brings the total text instances in our final dataset up to 79,814. Following the ICDAR15 standard, we annotated each image with all of the text instances, polygons, and content that were present. Manual annotations were done on each and every image.

SMDG

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.

Stairs Image Dataset | Parts of House | Indoor

This dataset is an extremely challenging set of over 3000+ originally Stair images captured and crowdsourced from over 500+ urban and rural areas, where each image is manually reviewed and verified by computer vision professionals at Datacluster Labs. ### Dataset Features - Dataset size : 3000+ - Captured by : Over 500+ crowdsource contributors - Resolution : 100% images HD and above (1920x1080 and above) - Location : Captured with 500+ cities accross India - Diversity : Various lighting conditions like day, night, varied distances, view points etc. - Device used : Captured using mobile phones in 2020-2022 - Usage : Stair detection , Stair Edge detection , Computer Vision , etc. ### Available Annotation formats COCO, YOLO, PASCAL-VOC, Tf-Record To download full datasets or to submit a request for your dataset needs, please ping us at sales@datacluster.ai Visit www.datacluster.ai to know more. Note: All the images are manually captured and verified by a large contributor base on DataCluster platform

TS-TR

The Turkish Scene Text Recognition (TS-TR) dataset was primarily developed to fill the gap in non-English text recognition resources, specifically addressing the unique challenges presented by the Turkish language, such as special characters and diacritics. This dataset mirrors real-world conditions with texts displayed in various fonts, sizes, orientations, and complex backgrounds from multiple urban and rural environments. Such diversity ensures the training of models that can generalize across different scenarios, including varying lighting conditions and complex visual layouts.

WTA/TLA

WTA (Wind Turbine Aerial) and TLA (Transmission Line Aerial) are public datasets which contain a set of RGB images from wind turbine farms and transmission towers and power lines, along with semantic ground truth for relevant classes. This is the official repository of the paper: WTA/TLA: A UAV-captured Dataset for Semantic Segmentation of Energy Infrastructure (url). Each dataset contains multiple locations. For each location: - images/ directory contains RGB images - multi_masks/ directory contains semantic image masks, with each class encoded in a different color (0:background, 1:blade, 2:tower). - annotations/ directory contains semantic image masks, for visual inspection. <h5>Training/Testing</h5> train.txt/test.txt and train_gt.txt/test_gt.txt contain images and masks for training and testing respectively <h5>Citation:</h5> @inproceedings{za2022wtatla, author={Zampokas, Georgios and Skartados, Evangelos and Alexiou, Dimitrios and Tsiakas, Kosmas and Tzanakis, Ioannis and Roussos, Nikolaos and Giakoumis, Dimitrios and Kostavelis, Ioannis and Bouganis, Christos-Savvas and Tzovaras, Dimitrios}, booktitle={2022 International Conference on Unmanned Aircraft Systems (ICUAS)}, title={WTA/TLA: A UAV-captured Dataset for Semantic Segmentation of Energy Infrastructure}, year={2022}

출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.