Papers with Code 아카이브

2D Object Detection

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
15
평가표
20
보존된 평가 기록
57

데이터셋

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

TBBR

The dataset of Thermal Bridges on Building Rooftops (TBBR dataset) consists of annotated combined RGB and thermal drone images with a height map. All images were converted to a uniform format of 3000$\times$4000 pixels, aligned, and cropped to 2400$\times$3400 to remove empty borders. The raw images for our dataset were recorded with a normal (RGB) and a FLIR-XT2 (thermal) camera on a DJI M600 drone. They show six large building blocks of around 20 buildings per block recorded in the city centre of the German city Karlsruhe east of the market square. Because of a high overlap rate of the images, the same buildings are on average recorded from different angles in different images about 20 times. All images were recorded during a drone flight on March 19, 2019 from 7 a.m. to 8 a.m. At this time, temperatures were between 3.78$^{\circ}$ C and 4.97$^{\circ}$ C, humidity between 80% and 98%. There was no rain on the day of the flight, but there was 2.3 $mm/m^2$ 48 hours beforehand. For recording the thermographic images an emissivity of 1.0 was set. The global radiation during this period was between 38.59 $W / m^2$ and 120.86 $W / m^2$. No direct sunlight can be seen visually on any of the recordings. The dataset contains 924 images with a total of 6930 annotations of thermal bridges on rooftops, split into train and test subsets with 722 (5614) and 202 (1313) images (annotations), respectively. The annotations only include thermal bridges that are visually identifiable with the human eye. Because of the aforementioned image overlap, each thermal bridge is annotated multiple times from different angles. For the annotation of the thermal images the image processing program VGG Image Annotator from the Visual Geometry Group, version 2.0.10, was used. The thermal bridge annotations are outlined with polygon shapes. These polygon lines were placed as close as possible but outside the area of significant temperature increase. If a detected thermal bridge was partially covered by another building component located in the foreground, the thermal bridge was also marked across the covering in case of minor coverings. Adjacent thermal bridges, which affect different rooftop components, were annotated separately. For example, a window with poor insulation of the window reveal located in the area of a poorly insulated roof is annotated individually. There is no overlap between annotated areas. While each image contains annotations, they also include thermal bridges present that are not annotated. Usage: Each compressed archive file represents one of the six building blocks. For the related publication the final block (Flug1_105Media) was used as a hold-out test sample. The archives contain Numpy files (one per image) of shape (2400, 3200, 5), where the final dimension is the color channel in the format [B, G, R, Thermal, Height]. Archives were compressed using ZStandard compression. They can be decompressed in a terminal by running e.g. tar -I zstd -xvf Flug1_105Media.tar.zst these will be decompressed into the file structure: images/ └── Flug1_105Media/ └── DJI_0004_R.npy └── DJI_0006_R.npy └── ... Corresponding annotations are provided in the COCO JSON format. There is one file for training (Flug1_100Media - Flug1_104Media blocks) and one for test (Flug1_105Media block). They contain a single class (thermal bridge) and expect the folder structure shown below. Note: The annotation files contain relative paths to numpy files, in case of problems please convert to absolute paths (i.e. insert the containing directory before each file path in the JSON annotation files). We recommend the following folder structure for reproduction of our work with Detectron2: ├── train/ │ ├── Flug1_100-104Media_coco.json │ └── images/ │ ├── Flug1_100Media/ │ │ ├── DJI_XXXX_R.npy │ │ └── ... │ ├── ... │ └── Flug1_104Media/ │ ├── DJI_XXXX_R.npy │ └── ... └── test/ ├── Flug1_105Media_coco.json └── images/ └── Flug1_105Media/ ├── DJI_XXXX_R.npy └── ...

ThermoHands

ThermoHands is the first benchmark dataset specifically designed for egocentric 3D hand pose estimation from thermal images. It addresses the challenges of hand pose estimation in low-light conditions and when the hand is occluded by gloves or other wearables—scenarios where traditional RGB or NIR-based systems struggle. The dataset contains approximately 96,000 synchronized multi-view, multi-spectral images collected from 28 participants under five varied scenarios involving changes in lighting, handwear, and background environments. Each sample includes thermal, NIR, RGB, and depth images, captured using a custom-built head-mounted sensor platform (HMSP) and a multi-view exocentric setup. Ground-truth 3D hand poses and shapes are automatically annotated using multi-view optimization based on the MANO model. Key features: Multi-spectral data: Egocentric and exocentric views of thermal, NIR, RGB, and depth modalities Rich annotations: 3D hand pose and shape ground truth via an automated optimization pipeline Realistic interactions: Both hand-object and hand-virtual interactions in diverse everyday scenarios Motivation and Use Cases: ThermoHands enables research into robust, privacy-preserving hand pose estimation for XR, AR/VR, and human-computer interaction applications. It is particularly useful for scenarios involving poor illumination, occlusion by gloves, or the need for passive sensing (e.g., avoiding NIR interference or power constraints). This dataset also introduces TherFormer, a novel transformer-based baseline tailored to thermal input, and benchmarks multiple SOTA methods under cross-spectral settings. Paper: https://arxiv.org/pdf/2403.09871 Project page: https://thermohands.github.io/ Code and models: https://github.com/LawrenceZ22/ThermoHands

TI1K Dataset

Thumb Index 1000 (TI1K) is a dataset of 1000 hand images with the hand bounding box, and thumb and index fingertip positions. The dataset includes the natural movement of the thumb and index fingers making it suitable for mixed reality (MR) applications. The dataset contains images only with the thumb and index fingers of both hands of resolution 640x480. All the annotations of the training and test images are in the "label.txt" file in the Annotation folder.

TNCR Dataset

We present TNCR, a new table dataset with varying image quality collected from free open source websites. TNCR dataset can be used for table detection in scanned document images and their classification into 5 different classes. TNCR contains 9428 high-quality labeled images. In this paper, we have implemented state-of-the-art deep learning-based methods for table detection to create several strong baselines. Cascade Mask R-CNN with ResNeXt-101-64x4d Backbone Network achieves the highest performance compared to other methods with a precision of 79.7%, recall of 89.8%, and f1 score of 84.4% on the TNCR dataset. We have made TNCR open source in the hope of encouraging more deep learning approaches to table detection, classification and structure recognition. Image source: https://github.com/abdoelsayed2016/TNCR_Dataset

Underwater Trash Detection

Underwater Trash Detection Dataset ## Overview The Underwater Trash Detection Dataset is a custom-annotated dataset designed to address the challenges of underwater trash detection caused by varying environmental features. Publicly available datasets alone are insufficient for training deep learning models due to domain-specific variations in underwater conditions. This dataset offers a cumulative, self-annotated collection of underwater images for detecting and classifying trash, providing a strong foundation for deep learning research and benchmark testing. --- ## Dataset Summary - Total Images: 9,576 - Annotation Types: Trash classification (plastic, trash, underwater debris) vs. environmental factors (fish, flora, fauna). --- ## Dataset Split | Split | Percentage | Number of Images | |-------------|----------------|-----------------------| | Train Set | 76% | 7,308 | | Validation | 19% | 1,795 | | Test Set | 5% | 473 | --- ## Preprocessing - Image Resize: All images are resized to 256x256 pixels using stretching for uniform input dimensions. --- ## Purpose This dataset supports research in underwater trash detection while addressing storage and computational constraints in underwater mobile devices. It enables the development of optimized algorithms for efficient trash detection and classification using minimal resources. --- ## Citation If you use this dataset in your research, please cite: ```bibtex @InProceedings{10.1007/978-3-031-43360-3_24, author="Walia, Jaskaran Singh and Seemakurthy, Karthik", editor="Iida, Fumiya and Maiolino, Perla and Abdulali, Arsen and Wang, Mingfeng", title="Optimized Custom Dataset for Efficient Detection of Underwater Trash", booktitle="Towards Autonomous Robotic Systems", year="2023", publisher="Springer Nature Switzerland", address="Cham", pages="292--303", }

xView3-SAR

Unsustainable fishing practices worldwide pose a major threat to marine resources and ecosystems. Identifying vessels that do not show up in conventional monitoring systems---known as ``dark vessels''---is key to managing and securing the health of marine environments. With the rise of satellite-based synthetic aperture radar (SAR) imaging and modern machine learning (ML), it is now possible to automate detection of dark vessels day or night, under all-weather conditions. SAR images, however, require a domain-specific treatment and are not widely accessible to the ML community. Maritime objects (vessels and offshore infrastructure) are relatively small and sparse, challenging traditional computer vision approaches. We present the largest labeled dataset for training ML models to detect and characterize vessels and ocean structures in SAR imagery. xView3-SAR consists of nearly 1,000 analysis-ready SAR images from the Sentinel-1 mission that are, on average, 29,400-by-24,400 pixels each. The images are annotated using a combination of automated and manual analysis. Co-located bathymetry and wind state rasters accompany every SAR image. We also provide an overview of the xView3 Computer Vision Challenge, an international competition using xView3-SAR for ship detection and characterization at large scale. We release the data (https://iuu.xview.us/) and code (https://github.com/DIUx-xView) to support ongoing development and evaluation of ML approaches for this important application.

((Refund~option))What is the refundable option on Expedia?

((+1-888-829-0881)) (US) or +1-888-829-0881 (UK): Expedia offers a refund option for certain reservations, allowing you to cancel and receive a full refund. To confirm eligibility, please refer to the Expedia hotel refund policy. If you have questions about the refund rate or cancellation fee, please contact an Expedia agent at ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK). Our agents can clarify your Expedia refund rate and provide you with more details. For immediate assistance, please call the Expedia customer service number with the area code: ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK). What is the refund option on Expedia? ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK): The Expedia refund option allows you to cancel specific reservations and receive a refund, depending on your hotel's refund policy. Please note that some reservations may have a refund or cancellation fee. For help with your options, please call ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK) to speak with an Expedia representative. You can also contact Expedia Customer Service with the area code: ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK) for more information. What is Expedia's refund option? ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK): If you need a refund option on Expedia, please check your hotel's refund policy to see if it qualifies for cancellation. Some reservations may have a refund fee, while others have a specific cancellation fee. For assistance, please contact an Expedia representative at ((+1-888-829-0881)) (US) or +1-888-829-0881 (UK). They will provide you with details about your refund rate and whether your reservation qualifies. Contact Expedia Customer Service with the area code: (+1-888-829-0881) (US) or +1-888-829-0881 (UK). What is the refund option on Expedia? (+1-888-829-0881) (US) or +1-888-829-0881 (UK): The Expedia refund option allows travelers to cancel specific reservations and receive a refund according to Expedia's hotel refund policy. It is important to review the terms of your reservation for any refund or cancellation fees. If you have any questions, please speak to an Expedia representative by calling (+1-888-829-0881) (US) or +1-888-829-0881 (UK). They can provide detailed explanations about your Expedia refund rate. For faster assistance, please contact Expedia Customer Service with the area code: (+1-888-829-0881) (US) or +1-888-829-0881 (UK). What is the Expedia refund option? (+1-888-829-0881) (US) or +1-888-829-0881 (UK): Expedia offers a refund option on certain reservations, but you should check your hotel's refund policy to determine if you qualify. Some cancellations may incur a refund charge or specific cancellation fee. If you're unsure of the terms, please contact an Expedia representative at (+1-888-829-0881) (US) or +1-888-829-0881 (UK) for assistance. They can clarify your Expedia refund fee and provide more details. For further assistance, please call Expedia Customer Service with the area code: (+1-888-829-0881) (US) or +1-888-829-0881 (UK).

3D-ZeF

3D-ZeF dataset consists of eight sequences with a duration between 15-120 seconds and 1-10 free moving zebrafish. The videos have been annotated with a total of 86,400 points and bounding boxes. Source: 3D-ZeF: A 3D Zebrafish Tracking Benchmark Dataset

[Travel@Guide® ]What is the 24 hour rule for KLM?

KLM's 24-hour rule allows passengers to cancel or modify their flight booking within 24 hours of purchase without penalty,☎+1||804||853||9001 or +1||801||855||5905✅ as long as the flight was booked at least 7 days before departure.

A collection of 131 CT datasets of pieces of modeling clay containing stones

This dataset contains a collection of 131 X-ray CT scans of pieces of modeling clay (Play-Doh) with various numbers of stones inserted, retrieved in the FleX-ray lab at CWI. The dataset consists of 5 parts. It is intended as raw supplementary material to reproduce the CT reconstructions and subsequent results in the paper titled "A tomographic workflow enabling deep learning for X-ray based foreign object detection". The dataset can be used to set up other CT-based experiments concerning similar objects with variations in shape and composition.

A Dataset of Multispectral Potato Plants Images

The dataset contains aerial agricultural images of a potato field with manual labels of healthy and stressed plant regions. The images were collected with a Parrot Sequoia multispectral camera carried by a 3DR Solo drone flying at an altitude of 3 meters. The dataset consists of RGB images with a resolution of 750×750 pixels, and spectral monochrome red, green, red-edge, and near-infrared images with a resolution of 416×416 pixels, and XML files with annotated bounding boxes of healthy and stressed potato crop.

A Large Scale Fish Dataset

This dataset contains 9 different seafood types collected from a supermarket in Izmir, Turkey for a university-industry collaboration project at Izmir University of Economics, and this work was published in ASYU 2020. The dataset includes gilt head bream, red sea bream, sea bass, red mullet, horse mackerel, black sea sprat, striped red mullet, trout, shrimp image samples. If you use this dataset in your work, please consider to cite: @inproceedings{ulucan2020large, title={A Large-Scale Dataset for Fish Segmentation and Classification}, author={Ulucan, Oguzhan and Karakaya, Diclehan and Turkan, Mehmet}, booktitle={2020 Innovations in Intelligent Systems and Applications Conference (ASYU)}, pages={1--5}, year={2020}, organization={IEEE} } The dataset contains 9 different seafood types. For each class, there are 1000 augmented images and their pair-wise augmented ground truths. Each class can be found in the "Fish_Dataset" file with their ground truth labels. All images for each class are ordered from "00000.png" to "01000.png". For example, if you want to access the ground truth images of the shrimp in the dataset, the order should be followed is "Fish->Shrimp->Shrimp GT". This dataset was collected in order to carry out segmentation, feature extraction, and classification tasks and compare the common segmentation, feature extraction, and classification algorithms (Semantic Segmentation, Convolutional Neural Networks, Bag of Features). All of the experiment results prove the usability of our dataset for purposes mentioned above.

Active Terahertz

This is a public dataset for evaluating multi-object detection algorithms in active Terahertz imaging resolution 5 mm by 5 mm.

AODRaw

We introduce the AODRaw dataset, which offers 7,785 high-resolution real RAW images with 135,601 annotated instances spanning 62 categories, capturing a broad range of indoor and outdoor scenes under 9 distinct light and weather conditions. AODRaw supports RAW and sRGB object detection.

Apron Dataset

The Apron Dataset focuses on training and evaluating classification and detection models for airport-apron logistics. In addition to bounding boxes and object categories the dataset is enriched with meta parameters to quantify the models’ robustness against environmental influences.

AquaTrash

This dataset contains 369 images of Trash used for deep learning. Each image is manually labelled by our team for accurate detections making a total of 470 bounding boxes. There are total 4 classes {(0: glass), (1:paper), (2:metal), (3:plastic)}

Aria Everyday Objects

A small-scale, real-world Project Aria dataset with high quality static 3D oriented bounding boxs annotations. Dataset Contents - Project Aria glasses data (including 2 x SLAM cameras, 1 x RGB camera, 2 x IMU, and complete sensor calibrations) - Aria machine perception service annotations including semi-dense point clouds and 6DoF device trajectory. - Manually-annotated 3D object bounding boxes Sequence Metrics - 45 minutes of egocentric recordings captured by non-computer vision experts across 25 diverse, primarily indoor, real-world environments - 1037 3D object bounding box instances across 17 classes: Bed, Chair, Couch, Door, Floor, Lamp, Mirror, Plant, Refrigerator, Screen, Sink, Storage, Table, Wall, WallArt, WasherDryer, and Window

BankNote-Net

Millions of people around the world have low or no vision. Assistive software applications have been developed for a variety of day-to-day tasks, including currency recognition. To aid with this task, we present BankNote-Net, an open dataset for assistive currency recognition. The dataset consists of a total of 24,816 embeddings of banknote images captured in a variety of assistive scenarios, spanning 17 currencies and 112 denominations. These compliant embeddings were learned using supervised contrastive learning and a MobileNetV2 architecture, and they can be used to train and test specialized downstream models for any currency, including those not covered by our dataset or for which only a few real images per denomination are available (few-shot learning). We deploy a variation of this model for public use in the last version of the Seeing AI app developed by Microsoft, which has over a 100 thousand monthly active users.

BBBC041

P. vivax (malaria) infected human blood smears with bounding box annotations. The data consists of two classes of uninfected cells (RBCs and leukocytes) and four classes of infected cells (gametocytes, rings, trophozoites, and schizonts).

Blood Cell Detection Dataset

Overview This is a dataset of blood cells photos. There are 364 images across three classes: WBC (white blood cells), RBC (red blood cells), and Platelets. There are 4888 labels across 3 classes (and 0 null examples). Fork this dataset (upper right hand corner) to receive the raw images, or (to save space) grab the 500x500 export. ## Use Cases This is a small scale object detection dataset, commonly used to assess model performance. It's a first example of medical imaging capabilities. This dataset is mainly preprocessed for YOLOV5 Application. ##Using this Dataset I'm releasing the data as public domain. Feel free to use it for any purpose. This dataset is already splitted into train,testing and validation datasets(70% for training, 20% testing and 10% for validation). The train,testing and validation folders are further classified as IMAGES AND LABELS. images Folder : It contains images of blood cells. labels Folder : It conatins labelling of blood cells across three classes: WBC (white blood cells), RBC (red blood cells), and Platelets. Except for GitHub, this dataset is also published on Kaggle.

Boreal Forest Fire

This dataset consists of annotated images and videos of smoke resulting from prescribed burning events in Finnish boreal forests. The dataset was created to train and validate learning-based methods for wildfire detection and smoke segmentation and its effectiveness in doing so was shown in the linked studies. The data was captured as 4K (3840×2160) videos at four events in Evo, Heinola, Karkkila, and Ruokolahti, using a DJI Phantom 4 drone. Individual frames of the videos were annotated manually with bounding boxes to enable the use of the data to train and test wildfire detection models. A portion of the bounding box annotated image data was resorted and annotated at the pixel level for image segmentation model training, validation, and testing. The training and validation data were annotated automatically using the Segment Anything Model and the manually annotated bounding boxes, while a small test set was annotated with manually drawn pixelwise masks. The three parts of the dataset are stored in the three separate directories: * Boreal-Forest-Fire-Subset-A: The bounding box annotated image data * Boreal-Forest-Fire-Subset-B: 30-second 4K video clips with binary annotations for smoke * Boreal-Forest-Fire-Subset-C: The segmentation mask annotated image data In addition, a Jupyter notebook providing an example for visualising and converting the bounding box coordinates is provided in the Visualisation-And-Coordinate-Conversion-Notebook directory. Code for using the segmentation data is found in the linked Gitlab repository.

Bramble flower image dataset

This dataset contains both the artificial and real flower images of bramble flowers. The real images were taken with a realsense D435 camera inside the West Virginia University greenhouse. All the flowers are annotated in YOLO format with bounding box and class name. The trained weights after training also have been provided. They can be used with the python script provided to detect the bramble flowers. Also the classifier can classify whether the flowers center is visible or hidden which will be helpful in precision pollination projects. Images are also augmented to make the task robust in various environmental conditions.

CAGUI

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

CAT: Context Adjustment Training

CAT is a specialized dataset for co-saliency detection. This dataset is intended for both helping to assess the performance of vision algorithms and supporting research that aims to exploit large volumes of annotated data, e.g., for training deep neural networks. Scale & Features - A total number of 33500 image samples. - 280 semantic groups affiliated to 15 superclasses. - High-quality mask annotations. - Diverse visual context with multiple foreground objects.

Cattle

Cattle data set, which was introduced in a paper. We (not the authors) created a train-val-test split.

CD-HARD

CD-HARD comprises 102 images featuring vehicles with oblique license plates sourced from the Cars dataset. Each image within this dataset exclusively depicts a single vehicle and was captured during daylight hours. While the dataset encompasses images from diverse geographic regions, it predominantly consists of images seemingly taken in European locales. The dataset was presented in the 'License Plate Detection and Recognition in Unconstrained Scenarios' research paper.

Charlotte-ThermalFace

Charlotte-ThermalFace is a thermal face dataset. The data is fully annotated with the facial landmarks, ambient temperature, relative humidity, the air speed of the room, distance to the camera, and subject thermal sensation at the time of capturing each image. There are approximately 10,000 infrared thermal images from 10 subjects in varying thermal conditions, at several distances from the camera, and at changing head positions. We have also controlled the air temperature to change from 20.5°C ( 69°F) to 26.5 °C( 80°F). Images are available in four different temperatures, 10 relative distances from the camera, starting at 1m ( 3.3 ft) to 6.6m( 21.6 ft), and 25 head positions. • The first public facial thermal dataset annotated with the environmental properties including air temperature, relative humidity, airspeed, distance from the camera, and subjective thermal sensation of each person at the time. • All the images are manually annotated with 72 or 43 facial landmarks. • We are publishing the data in the original 16-bit radiometric TemperatureLinear format, which has the thermal value of each pixel.

ChessReD2K

The Chess Recognition Dataset 2K (ChessReD2K) comprises a diverse collection of images of chess formations captured using smartphone cameras; a sensor choice made to ensure real-world applicability. The dataset is accompanied by detailed annotations providing information about the chess pieces formation in the images, bounding-boxes, and chessboard corner annotations. The number of annotations for each image depends on the number of chess pieces depicted in it. There are 12 category ids in total (i.e., 6 piece types per colour) and the chessboard coordinates are in the form of algebraic notation strings (e.g., "a8"). The corners are annotated based on their location on the chessboard (e.g., "bottom-left") with respect to the white player's view. This discrimination between these different types of corners provides information about the orientation of the chessboard that can be leveraged to determine the image's perspective and viewing angle. Dataset specifications The dataset consists of 20 chess games (selected from the ChessReD dataset), each with an arbitrary number of moves and therefore images, amounting to a total of 2,078 images. A 70/15/15 split stratified over the smartphone cameras was followed, which led to a total of 14 training games (1,442 images), 3 validation games (330 images), and 3 test games (306 images) being annotated. The split was also stratified over the three distinct smartphone cameras (Apple iPhone 12, Huawei P40 pro, Samsung Galaxy S8) that were used to capture the images.

CISOL

The Construction Industry Steel Ordering Lists (CISOL) dataset comprises table-centric, real-world documents from the construction industry, annotated to facilitate the testing and training of deep learning models for table detection (TD) and table structure recognition (TSR). Access to the CISOL Leaderboard is provided via https://eval.ai/web/challenges/challenge-page/2257

CLAD

CLAD (Compled and Long Activities Dataset) is an activity dataset which exhibits real-life and diverse scenarios of complex, temporally-extended human activities and actions. The dataset consists of a set of videos of actors performing everyday activities in a natural and unscripted manner. The dataset was recorded using a static Kinect 2 sensor which is commonly used on many robotic platforms. The dataset comprises of RGB-D images, point cloud data, automatically generated skeleton tracks in addition to crowdsourced annotations. Source: Complex and Long Activities Dataset

CLCXray

The CLCXray dataset contains 9,565 X-ray images, in which 4,543 X-ray images (real data) are obtained from the real subway scene and 5,022 X-ray images (simulated data) are scanned from manually designed baggage. There are 12 categories in the CLCXray dataset, including 5 types of cutters and 7 types of liquid containers. Five kinds of cutters include blade, dagger, knife, scissors, swiss army knife. Seven kinds of liquid containers include cans, carton drinks, glass bottle, plastic bottle, vacuum cup, spray cans, tin. The annotations are made in COCO format.

CLPD

The CLPD dataset comprises 1200 images that encompass various regions within mainland China. These images were sourced from diverse origins, including the internet, mobile devices, and in-car recording devices. While the majority of the images were recorded during daylight hours, a portion of them were captured at nighttime. The dataset predominantly features passenger cars, with a limited number of images depicting trucks and buses. The dataset was presented in the research paper titled "A Robust Attentional Framework for License Plate Recognition in the Wild."

COCO-Mix

COCO-Mixed dataset includes 897 images with annotations of both known and unknown categories. It contains 2533 unknown objects and 2658 known objects, with original COCO annotations used as labels for known objects. Unambiguous unlabeled objects are also annotated. The dataset is more challenging to evaluate due to the images containing more object instances with complex categories and concentrated locations.

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