Papers with Code 아카이브

Instance Segmentation

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
17
평가표
36
보존된 평가 기록
338

Instance Segmentation is a computer vision task that involves identifying and separating individual objects within an image, including detecting the boundaries of each object and assigning a unique label to each object. The goal of instance segmentation is to produce a pixel-wise segmentation map of the image, where each pixel is assigned to a specific object instance.

Image Credit: Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers, CVPR'21

데이터셋

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

ShipSG

The ShipSG dataset is the first public dataset of its kind for ship segmentation and georeferencing. The dataset has been made available for the development and evaluation of instance segmentation and georeferencing methods using computer vision and deep learning, thus advancing the research field of ship recognition for maritime situational awareness. ShipSG consists of 3505 images of a port location using two different cameras with static and oblique view. In total, 11625 ship masks grouped in seven ship classes were manually annotated. Moreover, each image contains one geographic annotation of one of the ship masks (latitude and longitude) present in the image.

BUP20

Video sequences from a glasshouse environment in Campus Kleinaltendorf(CKA), University of Bonn, captured by PATHoBot, a glasshouse monitoring robot. - 10 Video sequences, 120s long. - 2 cultivar Mavera (yellow) and All- rounder (red). - RGB-D images (Intel RealSense D435i cameras). - Robot odometry and IMU. - High quality sparese instance segmentation labels.

ISBDA

Consists of user-generated aerial videos from social media with annotations of instance-level building damage masks. This provides the first benchmark for quantitative evaluation of models to assess building damage using aerial videos. Source: MSNet: A Multilevel Instance Segmentation Network for Natural Disaster Damage Assessment in Aerial Videos

NERDS 360

We present a large-scale dataset for 3D urban scene understanding. Compared to existing datasets, our dataset consists of 75 outdoor urban scenes with diverse backgrounds, encompassing over 15,000 images. These scenes offer 360◦ hemispherical views, capturing diverse foreground objects illuminated under various lighting conditions. Additionally, our dataset encompasses scenes that are not limited to forward-driving views, addressing the limitations of previous datasets such as limited overlap and coverage between camera views. The closest pre-existing dataset for generalizable evaluation is DTU [2] (80 scenes) which comprises mostly indoor objects and does not provide multiple foreground objects or background scenes. We use the Parallel Domain synthetic data generation to render high-fidelity 360◦ scenes. We select 3 different maps i.e. SF 6thAndMission, SF GrantAndCalifornia and SF VanNessAveAndTurkSt and sample 75 different scenes across all 3 maps as our backgrounds (All 75 scenes across 3 maps are significantly different road scenes from each other, captured at different viewpoints in the city). We select 20 different cars in 50 different textures for training and randomly sample from 1 to 4 cars to render in a scene. We refer to this dataset as NeRDS 360: NeRF for Reconstruction, Decomposition and Scene Synthesis of 360◦ outdoor scenes. In total, we generate 15k renderings by sampling 200 cameras in a hemispherical dome at a fixed distance from the center of cars. We held out 5 scenes with 4 different cars and different backgrounds for testing, comprising 100 cameras distributed uniformly sampled in the upper hemisphere, different from the camera distributions used for training. We use the diverse validation camera distribution to test our approach’s ability to generalize to unseen viewpoints as well as unseen scenes during training. Our dataset and the corresponding task is extremely challenging due to occlusions, diversity of backgrounds, and rendered objects with various lightning and shadows. Our task entails reconstructing 360◦ hemispherical views of complete scenes using a handful of observations i.e. 1 to 5 as shown by red cameras whereas evaluating using all 100 hemispherical views, hence our task requires strong priors for novel view synthesis of outdoor scenes. Tasks our dataset support: - Generaliazable Novel view synthesis (Few shot evaluation) - Novel view synthesis (Overfitting evaluation) - 6D pose estimation - Object editing - Depth estimation - Semantic Segmentation - Instance Segmentation

ObjectsRoom

The ObjectsRoom dataset is based on the MuJoCo environment used by the Generative Query Network [4] and is a multi-object extension of the 3d-shapes dataset. The training set contains 1M scenes with up to three objects. We also provide ~1K test examples for the following variants: 2.1 Empty room: scenes consist of the sky, walls, and floor only. 2.2 Six objects: exactly 6 objects are visible in each image. 2.3 Identical color: 4-6 objects are placed in the room and have an identical, randomly sampled color. Datapoints consist of an image and fixed number of masks. The first four masks correspond to the sky, floor, and two halves of the wall respectively. The remaining masks correspond to the foreground objects. Source: Objects Room

OoDIS

OoDIS is a benchmark dataset for anomaly instance segmentation, crucial for autonomous vehicle safety. It extends existing anomaly segmentation benchmarks to focus on the segmentation of individual out-of-distribution (OOD) objects. The dataset addresses the need for identifying and segmenting unknown objects, which are critical to avoid accidents. It includes diverse scenes with various anomalies, pushing the boundaries of current segmentation capabilities. The benchmark is focused on evaluation of detection and instance segmentation of unexpected obstacles on roads. For more details, refer to the OoDIS paper

SpaceNet 1

SpaceNet 1: Building Detection v1 is a dataset for building footprint detection. The data is comprised of 382,534 building footprints, covering an area of 2,544 sq. km of 3/8 band WorldView-2 imagery (0.5 m pixel res.) across the city of Rio de Janeiro, Brazil. The images are processed as 200m×200m tiles with associated building footprint vectors for training. The main use case for the detection of building footprints from satellite imagery is to aid foundational mapping.

WGISD

Embrapa Wine Grape Instance Segmentation Dataset (WGISD) contains grape clusters properly annotated in 300 images and a novel annotation methodology for segmentation of complex objects in natural images. Source: Grape detection, segmentation and tracking using deep neural networks and three-dimensional association

ARMBench

ARMBench is a large-scale, object-centric benchmark dataset for robotic manipulation in the context of a warehouse. ARMBench contains images, videos, and metadata that corresponds to 235K+ pick-and-place activities on 190K+ unique objects. The data is captured at different stages of manipulation, i.e., pre-pick, during transfer, and after placement. Source: ARMBench: An Object-centric Benchmark Dataset for Robotic Manipulation Image Source: ARMBench: An Object-centric Benchmark Dataset for Robotic Manipulation

CREMI

MICCAI Challenge on Circuit Reconstruction from Electron Microscopy Images. # About The goal of this challenge is to evaluate algorithms for automatic reconstruction of neurons and neuronal connectivity from serial section electron microscopy data. The comparison is performed not only by evaluating the quality of neuron segmentations, but also by assessing the accuracy of detecting synapses and identifying synaptic partners. The challenge is carried out on three large and diverse datasets from adult Drosophila melanogaster brain tissue, comprising neuron segmentation ground truth and annotations for synaptic connections. A successful solution would demonstrate its efficiency and generalizability, and carry great potential to reduce the time spent on manual reconstruction of neural circuits in electron microscopy volumes. # Description We provide three datasets, each consisting of two (5 μm)3 volumes (training and testing, each 1250 px × 1250 px × 125 px) of serial section EM of the adult fly brain. Each volume has neuron and synapse labelings and annotations for pre- and post-synaptic partners.

Fraunhofer IPA Bin-Picking

The Fraunhofer IPA Bin-Picking dataset is a large-scale dataset comprising both simulated and real-world scenes for various objects (potentially having symmetries) and is fully annotated with 6D poses. A pyhsics simulation is used to create scenes of many parts in bulk by dropping objects in a random position and orientation above a bin. Additionally, this dataset extends the Siléane dataset by providing more samples. This allows to e.g. train deep neural networks and benchmark the performance on the public Siléane dataset Source: https://www.bin-picking.ai/en/dataset.html Image Source: https://arxiv.org/abs/1912.12125

iShape

iShape is an irregular shape dataset for instance segmentation. iShape contains six sub-datasets with one real and five synthetics, each represents a scene of a typical irregular shape.

MultiScan

We introduce MultiScan, a scalable RGBD dataset construction pipeline leveraging commodity mobile devices to scan indoor scenes with articulated objects and web-based semantic annotation interfaces to efficiently annotate object and part semantics and part mobility parameters. We use this pipeline to collect 273 scans of 117 indoor scenes containing 10957 objects and 5129 parts. The resulting MultiScan dataset provides RGBD streams with per-frame camera poses, textured 3D surface meshes, richly annotated part-level and object-level semantic labels, and part mobility parameters. We validate our dataset on instance segmentation and part mobility estimation tasks and benchmark methods for these tasks from prior work. Our experiments show that part segmentation and mobility estimation in real 3D scenes remain challenging despite recent progress in 3D object segmentation.

PeopleSansPeople

In recent years, person detection and human pose estimation have made great strides, helped by large-scale labeled datasets. However, these datasets had no guarantees or analysis of human activities, poses, or context diversity. Additionally, privacy, legal, safety, and ethical concerns may limit the ability to collect more human data. An emerging alternative to real-world data that alleviates some of these issues is synthetic data. However, creation of synthetic data generators is incredibly challenging and prevents researchers from exploring their usefulness. Therefore, we release a human-centric synthetic data generator PeopleSansPeople which contains simulation-ready 3D human assets, a parameterized lighting and camera system, and generates 2D and 3D bounding box, instance and semantic segmentation, and COCO pose labels. Using PeopleSansPeople, we performed benchmark synthetic data training using a Detectron2 Keypoint R-CNN variant [1]. We found that pre-training a network using synthetic data and fine-tuning on target real-world data (few-shot transfer to limited subsets of COCO-person train [2]) resulted in a keypoint AP of 60.37±0.48 (COCO test-dev2017) outperforming models trained with the same real data alone (keypoint AP of 55.80) and pre-trained with ImageNet (keypoint AP of 57.50). This freely-available data generator should enable a wide range of research into the emerging field of simulation to real transfer learning in the critical area of human-centric computer vision.

Separated COCO

Separated COCO is automatically generated subsets of COCO val dataset, collecting separated objects for a large variety of categories in real images in a scalable manner, where target object segmentation mask is separated into distinct regions by the occluder.

TexBiG

TexBiG (from the German Text-Bild-Gefüge, meaning Text-Image-Structure) is a document layout analysis dataset for historical documents in the late 19th and early 20th century. The dataset provides instance segmentation (bounding boxes and polygons/masks) annotations for 19 different classes with more then 52.000 instances. The added test images can be used to make submission on the leaderboard on EvalAI: https://eval.ai/web/challenges/challenge-page/2078/overview Dataset link: https://doi.org/10.5281/zenodo.8347059 Dataset (only train): https://www.kaggle.com/datasets/davidtschirschwitz/texbig-v2-0-train-val Dataset (only test image): https://www.kaggle.com/datasets/davidtschirschwitz/texbig-v2-0-test Please use TexBiG 2023 (which is v2.0 of the dataset) for testing model performance. The test dataset from the 2022 version (v1.0) are included as training data in v2.0 Each image of the dataset was annotated at least by two different annotators.

TimberSeg 1.0

The TimberSeg 1.0 dataset is composed of 220 images showing wood logs in various environments and conditions in Canada. The images are densely annotated with segmentation masks for each log instance, as well as the corresponding bounding box and class label. This dataset aim towards enabling autonomous forestry forwarders, therefore it contains nearly 2500 instances of wood logs from an operators' point-of-view. Images were taken in the forest, near the roadside, in lumberyards and above timber-filled trailers. The logs were annotated considering a grasping perspective, meaning that only the logs above the piles and accessible are segmented.

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} }

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.

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

GraSP

Holistic and Multi-Granular Surgical Scene Understanding of Prostatectomies (GraSP) dataset, a curated benchmark that models surgical scene understanding as a hierarchy of complementary tasks with varying levels of granularity. Our approach enables a multi-level comprehension of surgical activities, encompassing long-term tasks such as surgical phases and steps recognition and short-term tasks including surgical instrument segmentation and atomic visual actions detection. To exploit our proposed benchmark, we introduce the Transformers for Actions, Phases, Steps, and Instrument Segmentation (TAPIS) model, a general architecture that combines a global video feature extractor with localized region proposals from an instrument segmentation model to tackle the multi-granularity of our benchmark. Through extensive experimentation, we demonstrate the impact of including segmentation annotations in short-term recognition tasks, highlight the varying granularity requirements of each task, and establish TAPIS's superiority over previously proposed baselines and conventional CNN-based models. Additionally, we validate the robustness of our method across multiple public benchmarks, confirming the reliability and applicability of our dataset. This work represents a significant step forward in Endoscopic Vision, offering a novel and comprehensive framework for future research towards a holistic understanding of surgical procedures.

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.

Parcel2D Real

Real-world dataset of ~400 images of cuboid-shaped parcels with full 2D and 3D annotations in the COCO format.

RobotPush

RobotPush is a dataset for object singulation – the task of separating cluttered objects through physical interaction. The dataset contains 3456 training images with labels and 1024 validation images with labels. It consists of simulated and real-world data collected from a PR2 robot that equipped with a Kinect 2 camera. The dataset also contains ground truth instance segmentation masks for 110 images in the test set. Source: http://robotpush.cs.uni-freiburg.de/ Image Source: http://robotpush.cs.uni-freiburg.de/

UIIS10K

We propose a large-scale underwater instance segmentation dataset, UIIS10K, which includes 10,048 images with pixel-level annotations for 10 categories. As far as we know, this is the largest underwater instance segmentation dataset available and can be used as a benchmark for evaluating underwater segmentation methods.

Lindenthal Camera Traps

This data set contains 775 video sequences, captured in the wildlife park Lindenthal (Cologne, Germany) as part of the AMMOD project, using an Intel RealSense D435 stereo camera. In addition to color and infrared images, the D435 is able to infer the distance (or “depth”) to objects in the scene using stereo vision. Observed animals include various birds (at daytime) and mammals such as deer, goats, sheep, donkeys, and foxes (primarily at nighttime). A subset of 412 images is annotated with a total of 1038 individual animal annotations, including instance masks, bounding boxes, class labels, and corresponding track IDs to identify the same individual over the entire video. Source: Lindenthal Camera Traps on lila.science Image Source: Lindenthal Camera Traps on lila.science

LoTE-Animal

Understanding and analyzing animal behavior is increasingly essential to protect endangered animal species. However, the application of advanced computer vision techniques in this regard is minimal, which boils down to lacking large and diverse datasets for training deep models. To break the deadlock, we present LoTE-Animal, a large-scale endangered animal dataset collected over 12 years, to foster the application of deep learning in rare species conservation. The collected data contains vast variations such as ecological seasons, weather conditions, periods, viewpoints, and habitat scenes. So far, we retrieved at least 500K videos and 1.2 million images. Specifically, we selected and annotated 11 endangered animals for behavior understanding, including 10K video sequences for the action recognition task, 28K images for object detection, instance segmentation, and pose estimation tasks. In addition, we gathered 7K web images of the same species as source domain data for the domain adaptation task. We provide evaluation results of representative vision understanding approaches and cross-domain experiments. LoTE-Animal dataset would facilitate the community to research more advanced machine learning models and explore new tasks to aid endangered animal conservation. Our dataset will be released with the paper.

MOTFront

MOTFront provides photo-realistic RGB-D images with their corresponding instance segmentation masks, class labels, 2D & 3D bounding boxes, 3D geometry, 3D poses and camera parameters. The MOTFront dataset comprises 2,381 unique indoor sequences with a total of 60,000 images and is based on the 3D-FRONT dataset.

MVTec D2S

MVTec D2S is a benchmark for instance-aware semantic segmentation in an industrial domain. It contains 21,000 high-resolution images with pixel-wise labels of all object instances. The objects comprise groceries and everyday products from 60 categories. The benchmark is designed such that it resembles the real-world setting of an automatic checkout, inventory, or warehouse system. The training images only contain objects of a single class on a homogeneous background, while the validation and test sets are much more complex and diverse. Source: MVTec D2S: Densely Segmented Supermarket Dataset Image Source: https://www.mvtec.com/company/research/datasets/mvtec-d2s

Occluded COCO

Occluded COCO is automatically generated subset of COCO val dataset, collecting partially occluded objects for a large variety of categories in real images in a scalable manner, where target object is partially occluded but the segmentation mask is connected.

OmniCity

OmniCity is a dataset for omnipotent city understanding from multi-level and multi-view images. It contains multi-view satellite images as well as street-level panorama and mono-view images, constituting over 100K pixel-wise annotated images that are well-aligned and collected from 25K geo-locations in New York City. This dataset introduces a new task of fine-grained building instance segmentation on street-level panorama images. It also provides new problem settings for existing tasks, such as cross-view image matching, synthesis, segmentation, detection, etc., and facilitates the developing of new methods for large-scale city understanding, reconstruction, and simulation.

Parcel3D

Synthetic dataset of over 13,000 images of damaged and intact parcels with full 2D and 3D annotations in the COCO format. For details see our paper and for visual samples our project page.

Plittersdorf

A set of 221 stereo videos captured by the SOCRATES stereo camera trap in a wildlife park in Bonn, Germany between February and July of 2022. A subset of frames is labeled with instance annotations in the COCO format.

SB20

Video sequences captured at a field on Campus Kleinaltendorf (CKA), University of Bonn, captured by BonBot-I, an autonomous weeding robot. The data was captured by mounting an Intel RealSense D435i sensor with a nadir view of the ground. - RGB-D video sequences (Intel RealSense D435i cameras). - Robot odometry and IMU. - Crops and 8 different categories of weeds at different growth stages. - Different illumination conditions. - Three herbicide treatment regimes (30%, 70%, 100%), impacting weed density directly. - High quality sparese instance segmentation labels.

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