Papers with Code 아카이브

Semantic Segmentation

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
29
평가표
150
보존된 평가 기록
1,918

데이터셋

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

Endoscapes

Cholecystectomy is a very common abdominal surgical procedure almost ubiquitously performed with a laparoscopic approach, hence guided by an endoscopic video. Deep learning models for LC video analysis have been developed with the aim of assisting surgeons during interventions, improving staff awareness and readiness, and facilitating postoperative documentation and research. . However, datasets and models for video semantic segmentation of LC are lacking. Recognizing fine-grained hepatocystic anatomy through semantic segmentation could help surgeons better assess the critical view of safety (CVS), a universally recommended technique consisting in well exposing anatomical landmarks to prevent bile duct injuries. Additionally, segmentation masks of hepatocystic structures could be leveraged by deep learning models for automatic assessment of CVS and surgical action recognition to improve their performance. We believe that generating a dataset for video semantic segmentation of hepatocystic anatomy will promote surgical data science research and accelerate the development of applications for surgical safety. To generate a representative dataset, consecutive endoscopic videos of LC performed at Nouvel Hopital Civil (Strasbourg, France) were collected. Non-endoscopic, i.e., out-of-body, video frames were blackedout to comply with European data protection regulations. A frame every 30 seconds was sampled from the portion of the endoscopic video showing the hepatocystic anatomy being dissected, the most critical phase of the surgical procedure, and when surgeons should achieve the CVS. Such unselected and regularly spaced video frames were manually annotated with pixel-wise semantic annotations of anatomical and surgical instances, such as the cystic artery and the dissection. Overall, 1933 regularly spaced video frames from 201 LC videos were annotated with segmentation mask for 29 classes of the hepatocystic triangle, respectively. performed in double by specifically trained computer scientists and surgeons.

Fine-Grained Cloud Segmentation Dataset

The dataset consists of 96 terrain-corrected (Level-1T) scenes from Landsat 8 OLI and TIRS, covering diverse biomes. This variety supports cloud detection and removal in complex environments. The dataset includes manually generated cloud masks with pixel-level annotations for cloud shadow, clear sky, thin clouds, and cloud areas. Each scene is cropped into 512×512 pixel patches and split into training, validation, and test sets (6:2:2 ratio). It is a valuable resource for training and evaluating fine-grained cloud segmentation models across various terrains.

FLAIR (French Land cover from Aerospace ImageRy)

The French National Institute of Geographical and Forest Information (IGN) has the mission to document and measure land-cover on French territory and provides referential geographical datasets, including high-resolution aerial images and topographic maps. The monitoring of land-cover plays a crucial role in land management and planning initiatives, which can have significant socio-economic and environmental impact. Together with remote sensing technologies, artificial intelligence (IA) promises to become a powerful tool in determining land-cover and its evolution. IGN is currently exploring the potential of IA in the production of high-resolution land cover maps. Notably, deep learning methods are employed to obtain a semantic segmentation of aerial images. However, territories as large as France imply heterogeneous contexts: variations in landscapes and image acquisition make it challenging to provide uniform, reliable and accurate results across all of France. The FLAIR-one dataset presented is part of the dataset currently used at IGN to establish the French national reference land cover map "Occupation du sol `a grande 'echelle" (OCS- GE). It covers 810 km² and has 13 semantic classes.

Khanhha's dataset

From my knowledge, the dataset used in the project is the largest crack segmentation dataset so far. It contains around 11.200 images that are merged from 12 available crack segmentation datasets. The name prefix of each image is assigned to the corresponding dataset name that the image belong to. There're also images with no crack pixel, which could be filtered out by the file name pattern "noncrack*" All the images are resized to the size of (448, 448). The two folders images and masks contain all the images. The two folders train and test contain training and testing images splitted from the two above folder. The splitting is stratified so that the proportion of each dataset in the train and test folder are similar. Citation Note: please cite the corresponding papers when using these datasets. CRACK500: @inproceedings{zhang2016road, title={Road crack detection using deep convolutional neural network}, author={Zhang, Lei and Yang, Fan and Zhang, Yimin Daniel and Zhu, Ying Julie}, booktitle={Image Processing (ICIP), 2016 IEEE International Conference on}, pages={3708--3712}, year={2016}, organization={IEEE} }' . @article{yang2019feature, title={Feature Pyramid and Hierarchical Boosting Network for Pavement Crack Detection}, author={Yang, Fan and Zhang, Lei and Yu, Sijia and Prokhorov, Danil and Mei, Xue and Ling, Haibin}, journal={arXiv preprint arXiv:1901.06340}, year={2019} } GAPs384: @inproceedings{eisenbach2017how, title={How to Get Pavement Distress Detection Ready for Deep Learning? A Systematic Approach.}, author={Eisenbach, Markus and Stricker, Ronny and Seichter, Daniel and Amende, Karl and Debes, Klaus and Sesselmann, Maximilian and Ebersbach, Dirk and Stoeckert, Ulrike and Gross, Horst-Michael}, booktitle={International Joint Conference on Neural Networks (IJCNN)}, pages={2039--2047}, year={2017} } CFD: @article{shi2016automatic, title={Automatic road crack detection using random structured forests}, author={Shi, Yong and Cui, Limeng and Qi, Zhiquan and Meng, Fan and Chen, Zhensong}, journal={IEEE Transactions on Intelligent Transportation Systems}, volume={17}, number={12}, pages={3434--3445}, year={2016}, publisher={IEEE} } AEL: @article{amhaz2016automatic, title={Automatic Crack Detection on Two-Dimensional Pavement Images: An Algorithm Based on Minimal Path Selection.}, author={Amhaz, Rabih and Chambon, Sylvie and Idier, J{'e}r{^o}me and Baltazart, Vincent} } cracktree200: @article{zou2012cracktree, title={CrackTree: Automatic crack detection from pavement images}, author={Zou, Qin and Cao, Yu and Li, Qingquan and Mao, Qingzhou and Wang, Song}, journal={Pattern Recognition Letters}, volume={33}, number={3}, pages={227--238}, year={2012}, publisher={Elsevier} } https://github.com/alexdonchuk/cracks_segmentation_dataset https://github.com/yhlleo/DeepCrack https://github.com/ccny-ros-pkg/concreteIn_inpection_VGGF (Citing from https://github.com/khanhha/crack_segmentation.)

MCubeS (P)

Multimodal material segmentation (MCubeS) dataset contains 500 sets of images from 42 street scenes. Each scene has images for four modalities: RGB, angle of linear polarization (AoLP), degree of linear polarization (DoLP), and near-infrared (NIR). The dataset provides annotated ground truth labels for both material and semantic segmentation for every pixel. The dataset is divided training set with 302 image sets, validation set with 96 image sets, and test set with 102 image sets. Each image has 1224 x 1024 pixels and a total of 20 class labels per pixel.

MUAD

The MUAD dataset (Multiple Uncertainties for Autonomous Driving), consisting of 10,413 realistic synthetic images with diverse adverse weather conditions (night, fog, rain, snow), out-of-distribution objects, and annotations for semantic segmentation, depth estimation, object, and instance detection. Predictive uncertainty estimation is essential for the safe deployment of Deep Neural Networks in real-world autonomous systems and MUAD allows to a better assess the impact of different sources of uncertainty on model performance.

MUSES: MUlti-SEnsor Semantic perception dataset

MUSES offers 2500 multi-modal scenes, evenly distributed across various combinations of weather conditions (clear, fog, rain, and snow) and types of illumination (daytime, nighttime). Each image includes high-quality 2D pixel-level panoptic annotations and class-level and novel instance-level uncertainty annotations. Further, each adverse-condition image has a corresponding image of the same scene taken under clear-weather, daytime conditions. The annotation process for MUSES utilizes all available sensor data, allowing the annotators to also reliably label degraded image regions that are still discernible in other modalities. This results in better pixel coverage in the annotations and creates a more challenging evaluation setup. The dataset provides public benchmarks for: - Panoptic segmentation - Uncertainty-aware panoptic segmentation - Semantic segmentation - Object detection Sensor modalities included: - Frame camera (RGB) - MEMS lidar - FMCW radar - HD event camera - IMU/GNSS sensor

SWINySEG

The SWINySEG dataset contains 6768 daytime- and nighttime-images of sky/cloud patches along with their corresponding binary ground truth maps. The images in the SWINySeg dataset are taken from two of our earlier sky/cloud image segmentation datasets -- SWIMSEG and SWINSEG. All images were captured in Singapore using WAHRSIS, a calibrated ground-based whole sky imager, over a period of 12 months from January to December 2016. The ground truth annotation was done in consultation with experts from Singapore Meteorological Services. Source: CloudSegNet: A deep network for nychthemeron cloud image segmentation

Swiss3DCities

Swiss3DCities is a dataset that is manually annotated for semantic segmentation with per-point labels, and is built using photogrammetry from images acquired by multirotors equipped with high-resolution cameras. Source: Semantic Segmentation on Swiss3DCities: A Benchmark Study on Aerial Photogrammetric 3D Pointcloud Dataset Image Source: Can et al

USIS10K

We construct the first large-scale dataset, USIS10K, for the underwater salient instance segmentation task, which contains 10,632 images and pixel-level annotations of 7 categories. As far as we know, this is the largest salient instance segmentation dataset, and includes Class-Agnostic and Multi-Class labels simultaneously. If you use this dataset in your work, please consider to cite: @inproceedings{ lian2024icml, title={Diving into Underwater: Segment Anything Model Guided Underwater Salient Instance Segmentation and A Large-scale Dataset}, author={Shijie Lian and Ziyi Zhang and Hua Li and Wenjie Li and Laurence Tianruo Yang and Sam Kwong and Runmin Cong}, booktitle={Forty-first International Conference on Machine Learning}, year={2024}, url={https://openreview.net/forum?id=snhurpZt63} }

VLOG Dataset

A large collection of interaction-rich video data which are annotated and analyzed. Source: From Lifestyle Vlogs to Everyday Interactions

WISDOM

Synthetic training dataset of 50,000 depth images and 320,000 object masks using simulated heaps of 3D CAD models. Source: Segmenting Unknown 3D Objects from Real Depth Images using Mask R-CNN Trained on Synthetic Data

Aircraft Context Dataset

The Aircraft Context Dataset, a composition of two inter-compatible large-scale and versatile image datasets focusing on manned aircraft and UAVs, is intended for training and evaluating classification, detection and segmentation models in aerial domains. Additionally, a set of relevant meta-parameters can be used to quantify dataset variability as well as the impact of environmental conditions on model performance.

ARCH2S

Precise segmentation of architectural structures provides detailed information about various building components, enhancing our understanding and interaction with our built environment. Nevertheless, existing outdoor 3D point cloud datasets have limited and detailed annotations on architectural exteriors due to privacy concerns and the expensive costs of data acquisition and annotation. To overcome this shortfall, this paper introduces a semantically-enriched, photo-realistic 3D architectural models dataset and benchmark for semantic segmentation. It features 4 different building purposes of real-world buildings as well as an open architectural landscape in Hong Kong. Each point cloud is annotated into one of 14 semantic classes.

BRIGHT

BRIGHT is the first open-access, globally distributed, event-diverse multimodal dataset specifically curated to support AI-based disaster response. It covers five types of natural disasters and two types of man-made disasters across 14 disaster events in 23 regions worldwide, with a particular focus on developing countries. It supports not only the development of supervised deep models, but also the testing of their performance on cross-event transfer setup, as well as unsupervised domain adaptation, semi-supervised learning, unsupervised change detection, and unsupervised image matching methods in multimodal and disaster scenarios.

CheXlocalize

CheXlocalize is a radiologist-annotated segmentation dataset on chest X-rays. The dataset consists of two types of radiologist annotations for the localization of 10 pathologies: pixel-level segmentations and most-representative points. Annotations were drawn on images from the CheXpert validation and test sets. The dataset also consists of two separate sets of radiologist annotations: (1) ground-truth pixel-level segmentations on the validation and test sets, drawn by two board-certified radiologists, and (2) benchmark pixel-level segmentations and most-representative points on the test set, drawn by a separate group of three board-certified radiologists. The validation and test sets consist of 234 chest X-rays from 200 patients and 668 chest X-rays from 500 patients, respectively. The 10 pathologies of interest are Atelectasis, Cardiomegaly, Consolidation, Edema, Enlarged Cardiomediastinum, Lung Lesion, Lung Opacity, Pleural Effusion, Pneumothorax, and Support Devices. For more details, please see our paper, Benchmarking saliency methods for chest X-ray interpretation.

Cluttered Omniglot

Dataset for one-shot segmentation. Source: One-Shot Segmentation in Clutter

DeepSportRadar-v1

DeepSportradar is a benchmark suite of computer vision tasks, datasets and benchmarks for automated sport understanding. DeepSportradar currently supports four challenging tasks related to basketball: ball 3D localization, camera calibration, player instance segmentation and player re-identification. For each of the four tasks, a detailed description of the dataset, objective, performance metrics, and the proposed baseline method are provided. Source: DeepSportradar-v1: Computer Vision Dataset for Sports Understanding with High Quality Annotations Image Source: https://github.com/deepsportradar/instance-segmentation-challenge

DOORS

DOORS is a dataset designed for boulders recognition, centroid regression, segmentation, and navigation applications. The dataset is divided into two sets: - Regression: Contains images, masks, and labels for 4 splits of single boulders positioned on the surface of a spherical mesh. It can be used to perform navigation, boulder recognition, segmentation, and centroid regression. - Segmentation: Contain images, masks, and labels of 2 datasets: DS1 and DS2. DS1 is made of the same images of the Regression dataset but is specifically designed for segmentation. DS2 is made of images with multiple instances of boulders appearing on the surface of the Didymos asteroid model Source: DOORS: Dataset fOr bOuldeRs Segmentation. Statistical properties and Blender setup Image Source: https://arxiv.org/pdf/2210.16253v1.pdf

Five-Billion-Pixels

The Five-Billion-Pixels dataset contains more than 5 billion labeled pixels of 150 high-resolution Gaofen-2 (4 m) satellite images, annotated in a 24-category system covering artificial-constructed, agricultural, and natural classes. It possesses the advantage of rich categories, large coverage, wide distribution, and high-spatial resolution, which well reflects the distributions of real-world ground objects and can benefit to different land cover related studies.

HelixNet

Large-scale and open-access LiDAR dataset intended for the evaluation of real-time semantic segmentation algorithms. In contrast to other large-scale datasets, HelixNet includes fine-grained data about the sensor's rotation and position, as well as the points' release time.

Hephaestus

Hephaestus is the first large-scale InSAR dataset. Driven by volcanic unrest detection, it provides 19,919 unique satellite frames annotated with a diverse set of labels. Moreover, each sample is accompanied by a textual description of its contents. The goal of this dataset is to boost research on exploitation of interferometric data enabling the application of state-of-the-art computer vision+NLP methods. Furthermore, the annotated dataset is bundled with a large archive of unlabeled frames to enable large-scale self-supervised learning. The final size of the dataset amounts to 110,573 interferograms.

HSI-Drive v2.0

HSI-Drive is the hyperspectral image (HSI) dataset created by the Digital Electronics Design Group (GDED) of the University of the Basque Country (UPV/EHU). This database is intended to contribute to the research into the use of hyperspectral imaging for the development of advanced driver assistance systems (ADAS) and autonomous driving systems (ADS). The dataset contains a diverse set of images recorded with a small-size 25-band VNIR snapshot camera mounted on a moving automobile. The recordings have been made in different seasons of the year, at different day times, under different weather conditions and on different types of roads. The dataset contains images and videos classified and tagged accordingly to provide rich and diverse data.

HyKo2-VIS

We present datasets containing urban traffic and rural road scenes recorded using hyperspectral snap-shot sensors mounted on a moving car. The novel hyperspectral cameras used can capture whole spectral cubes at up to 15 Hz. This emerging new sensor modality enables hyperspectral scene analysis for autonomous driving tasks. Up to the best of the author’s knowledge no such dataset has been published so far. The datasets contain synchronized 3-D laser, spectrometer and hyperspectral data. Dense ground truth annotations are provided as semantic labels, material and traversability. The hyperspectral data ranges from visible to near infrared wavelengths. We explain our recoding platform and method, the associated data format along with a code library for easy data consumption. The datasets are publicly available for download.

Image and Video Advertisements

The Image and Video Advertisements collection consists of an image dataset of 64,832 image ads, and a video dataset of 3,477 ads. The data contains rich annotations encompassing the topic and sentiment of the ads, questions and answers describing what actions the viewer is prompted to take and the reasoning that the ad presents to persuade the viewer ("What should I do according to this ad, and why should I do it? "), and symbolic references ads make (e.g. a dove symbolizes peace). Source: Automatic Understanding of Image and Video Advertisements

Medico automatic polyp segmentation challenge (dataset)

The “Medico automatic polyp segmentation challenge” aims to develop computer-aided diagnosis systems for automatic polyp segmentation to detect all types of polyps (for example, irregular polyp, smaller or flat polyps) with high efficiency and accuracy. The main goal of the challenge is to benchmark semantic segmentation algorithms on a publicly available dataset, emphasizing robustness, speed, and generalization. Medico Multimedia Task at MediaEval 2020:Automatic Polyp Segmentation (https://arxiv.org/pdf/2012.15244.pdf)

MM-OR

Operating rooms (ORs) are complex, high-stakes environments requiring precise understanding of interactions among medical staff, tools, and equipment for enhancing surgical assistance, situational awareness, and patient safety. Current datasets fall short in scale, realism and do not capture the multimodal nature of OR scenes, limiting progress in OR modeling. To this end, we introduce MM-OR, a realistic and large-scale multimodal spatiotemporal OR dataset, and the first dataset to enable multimodal scene graph generation. MM-OR captures comprehensive OR scenes containing RGB-D data, detail views, audio, speech transcripts, robotic logs, and tracking data and is annotated with panoptic segmentations, semantic scene graphs, and downstream task labels. Further, we propose MM2SG, the first multimodal large vision-language model for scene graph generation, and through extensive experiments, demonstrate its ability to effectively leverage multimodal inputs. Together, MM-OR and MM2SG establish a new benchmark for holistic OR understanding, and open the path towards multimodal scene analysis in complex, high-stakes environments. Paper: https://arxiv.org/abs/2503.02579

Munich Sentinel2 Crop Segmentation

Contains squared blocks of 48×48 pixels including 13 Sentinel-2 bands. Each 480-m block was mined from a large geographical area of interest (102 km × 42 km) located north of Munich, Germany.

OAM-TCD

OAM-TCD is a dataset of around 5k aerial images from around the world to support robust tree detection algorithms. Full details can be found on the linked HuggingFace repository. Accurately quantifying tree cover is an important metric for ecosystem monitoring and for assessing progress in restored sites. Recent works have shown that deep learning-based segmentation algorithms are capable of accurately mapping trees at country and continental scales using high-resolution aerial and satellite imagery. Mapping at high (ideally sub-meter) resolution is necessary to identify individual trees, however there are few open-access datasets containing instance level annotations and those that exist are small or not geographically diverse. We present a novel open-access dataset for individual tree crown delineation (TCD) in high-resolution aerial imagery sourced from OpenAerialMap (OAM). Our dataset, OAM-TCD, comprises 5072 2048x2048 px images at 10 cm/px resolution with associated human-labeled instance masks for over 280k individual and 56k groups of trees. By sampling imagery from around the world, we are able to better capture the diversity and morphology of trees in different terrestrial biomes and in both urban and natural environments. Using our dataset, we train reference instance and semantic segmentation models that compare favorably to existing state-of-the-art models. We assess performance through k-fold cross-validation and comparison with existing datasets; additionally we demonstrate compelling results on independent aerial imagery captured over Switzerland and compare to municipal tree inventories and LIDAR-derived canopy maps in the city of Zurich. Our dataset, models and training/benchmark code are publicly released under permissive open-source licenses: Creative Commons (majority CC BY 4.0), and Apache 2.0 respectively.

OpenTrench3D

OpenTrench3D, the first publicly available point cloud dataset of underground utilities from open trenches. It features 310 fully annotated point clouds consisting of a total of 528 million points categorised into 5 unique classes. OpenTrench3D consists of photogrammetrically derived 3D point clouds capturing detailed scenes of open trenches, revealing underground utilities.

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