Ultralytics-中文文档-五-
Ultralytics 中文文档(五)
Argoverse 数据集
Argoverse数据集是由 Argo AI 开发的数据集,旨在支持自动驾驶任务的研究,如 3D 跟踪、运动预测和立体深度估计。该数据集提供多种高质量传感器数据,包括高分辨率图像、LiDAR 点云和地图数据。
注意
由于福特关闭 Argo AI 后,用于训练的 Argoverse 数据集*.zip
文件已从 Amazon S3 中删除,但我们已在Google Drive上提供手动下载。
主要特性
-
Argoverse 包含超过 290K 个标记的 3D 对象轨迹和 1263 个不同场景中的 500 万个对象实例。
-
数据集包括高分辨率相机图像、LiDAR 点云和丰富的 HD 地图标注。
-
标注包括对象的 3D 边界框、对象轨迹和轨迹信息。
-
Argoverse 为不同任务提供多个子集,如 3D 跟踪、运动预测和立体深度估计。
数据集结构
Argoverse 数据集分为三个主要子集:
-
Argoverse 3D 跟踪:该子集包含 113 个场景,超过 290K 个标记的 3D 对象轨迹,专注于 3D 对象跟踪任务。包括 LiDAR 点云、相机图像和传感器校准信息。
-
Argoverse 运动预测:该子集包含来自 60 小时驾驶数据的 324K 车辆轨迹,适用于运动预测任务。
-
Argoverse 立体深度估计:该子集专为立体深度估计任务设计,包括超过 10K 个立体图像对及相应的 LiDAR 点云,用于地面真实深度估计。
应用
Argoverse 数据集广泛用于训练和评估深度学习模型,用于自动驾驶任务,如 3D 对象跟踪、运动预测和立体深度估计。该数据集多样的传感器数据、对象标注和地图信息使其成为自动驾驶领域研究人员和从业者的宝贵资源。
数据集 YAML
使用 YAML(又一种标记语言)文件来定义数据集配置。它包含关于数据集路径、类别和其他相关信息。对于 Argoverse 数据集,Argoverse.yaml
文件维护在github.com/ultralytics/ultralytics/blob/main/ultralytics/cfg/datasets/Argoverse.yaml
。
ultralytics/cfg/datasets/Argoverse.yaml
# Ultralytics YOLO 🚀, AGPL-3.0 license
# Argoverse-HD dataset (ring-front-center camera) https://www.cs.cmu.edu/~mengtial/proj/streaming/ by Argo AI
# Documentation: https://docs.ultralytics.com/datasets/detect/argoverse/
# Example usage: yolo train data=Argoverse.yaml
# parent
# ├── ultralytics
# └── datasets
# └── Argoverse ← downloads here (31.5 GB)
# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
path: ../datasets/Argoverse # dataset root dir
train: Argoverse-1.1/images/train/ # train images (relative to 'path') 39384 images
val: Argoverse-1.1/images/val/ # val images (relative to 'path') 15062 images
test: Argoverse-1.1/images/test/ # test images (optional) https://eval.ai/web/challenges/challenge-page/800/overview
# Classes
names:
0: person
1: bicycle
2: car
3: motorcycle
4: bus
5: truck
6: traffic_light
7: stop_sign
# Download script/URL (optional) ---------------------------------------------------------------------------------------
download: |
import json
from tqdm import tqdm
from ultralytics.utils.downloads import download
from pathlib import Path
def argoverse2yolo(set):
labels = {}
a = json.load(open(set, "rb"))
for annot in tqdm(a['annotations'], desc=f"Converting {set} to YOLOv5 format..."):
img_id = annot['image_id']
img_name = a['images'][img_id]['name']
img_label_name = f'{img_name[:-3]}txt'
cls = annot['category_id'] # instance class id
x_center, y_center, width, height = annot['bbox']
x_center = (x_center + width / 2) / 1920.0 # offset and scale
y_center = (y_center + height / 2) / 1200.0 # offset and scale
width /= 1920.0 # scale
height /= 1200.0 # scale
img_dir = set.parents[2] / 'Argoverse-1.1' / 'labels' / a['seq_dirs'][a['images'][annot['image_id']]['sid']]
if not img_dir.exists():
img_dir.mkdir(parents=True, exist_ok=True)
k = str(img_dir / img_label_name)
if k not in labels:
labels[k] = []
labels[k].append(f"{cls} {x_center} {y_center} {width} {height}\n")
for k in labels:
with open(k, "w") as f:
f.writelines(labels[k])
# Download 'https://argoverse-hd.s3.us-east-2.amazonaws.com/Argoverse-HD-Full.zip' (deprecated S3 link)
dir = Path(yaml['path']) # dataset root dir
urls = ['https://drive.google.com/file/d/1st9qW3BeIwQsnR0t8mRpvbsSWIo16ACi/view?usp=drive_link']
print("\n\nWARNING: Argoverse dataset MUST be downloaded manually, autodownload will NOT work.")
print(f"WARNING: Manually download Argoverse dataset '{urls[0]}' to '{dir}' and re-run your command.\n\n")
# download(urls, dir=dir)
# Convert
annotations_dir = 'Argoverse-HD/annotations/'
(dir / 'Argoverse-1.1' / 'tracking').rename(dir / 'Argoverse-1.1' / 'images') # rename 'tracking' to 'images'
for d in "train.json", "val.json":
argoverse2yolo(dir / annotations_dir / d) # convert Argoverse annotations to YOLO labels
使用方法
要在 Argoverse 数据集上使用 YOLOv8n 模型进行 100 个 epoch 的训练,图像大小为 640,请使用以下代码片段。有关可用参数的全面列表,请参阅模型训练页面。
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="Argoverse.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=Argoverse.yaml model=yolov8n.pt epochs=100 imgsz=640
样本数据和注释
Argoverse 数据集包含各种传感器数据,包括摄像头图像、LiDAR 点云和高清地图信息,为自动驾驶任务提供丰富的背景信息。以下是数据集中的一些示例数据及其对应的注释:
- Argoverse 3D 跟踪:此图展示了 3D 物体跟踪的示例,物体用 3D 边界框进行了注释。数据集提供 LiDAR 点云和摄像头图像,以促进为此任务开发模型。
该示例展示了 Argoverse 数据集中数据的多样性和复杂性,并突显了高质量传感器数据在自动驾驶任务中的重要性。
引用和致谢
如果您在研究或开发工作中使用 Argoverse 数据集,请引用以下论文:
@inproceedings{chang2019argoverse,
title={Argoverse: 3D Tracking and Forecasting with Rich Maps},
author={Chang, Ming-Fang and Lambert, John and Sangkloy, Patsorn and Singh, Jagjeet and Bak, Slawomir and Hartnett, Andrew and Wang, Dequan and Carr, Peter and Lucey, Simon and Ramanan, Deva and others},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={8748--8757},
year={2019}
}
我们要感谢 Argo AI 创建和维护 Argoverse 数据集,作为自动驾驶研究社区的宝贵资源。有关 Argoverse 数据集及其创建者的更多信息,请访问Argoverse 数据集网站。
常见问题解答
什么是 Argoverse 数据集及其主要特点?
由 Argo AI 开发的Argoverse数据集支持自动驾驶研究。它包括超过 290K 个标记的 3D 物体轨迹和 1,263 个独特场景中的 5 百万个物体实例。数据集提供高分辨率摄像头图像、LiDAR 点云和标记的高清地图,对于 3D 跟踪、运动预测和立体深度估计等任务非常有价值。
如何使用 Argoverse 数据集训练 Ultralytics YOLO 模型?
要使用 Argoverse 数据集训练 YOLOv8 模型,请使用提供的 YAML 配置文件和以下代码:
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="Argoverse.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=Argoverse.yaml model=yolov8n.pt epochs=100 imgsz=640
有关参数详细说明,请参考模型训练页面。
Argoverse 数据集中提供了哪些数据类型和注释?
Argoverse 数据集包括各种传感器数据类型,如高分辨率摄像头图像、LiDAR 点云和高清地图数据。注释包括 3D 边界框、物体轨迹和轨迹信息。这些全面的注释对于准确地在 3D 物体跟踪、运动预测和立体深度估计等任务中进行模型训练至关重要。
Argoverse 数据集的结构是如何的?
数据集分为三个主要子集:
-
Argoverse 3D 跟踪:包括 113 个场景,超过 290K 个标记的 3D 物体轨迹,重点关注 3D 物体跟踪任务。它包括 LiDAR 点云、摄像头图像和传感器校准信息。
-
Argoverse 运动预测:包括从 60 小时驾驶数据中收集的 324K 车辆轨迹,适用于运动预测任务。
-
Argoverse 立体深度估计:包含超过 10K 对立体图像和相应的 LiDAR 点云,用于地面真实深度估计。
Argoverse 数据集已从 Amazon S3 中移除,我现在从哪里可以下载?
之前在 Amazon S3 上可用的 Argoverse 数据集*.zip
文件现在可以从Google Drive手动下载。
Argoverse 数据集中的 YAML 配置文件用于什么目的?
一个 YAML 文件包含数据集的路径、类别和其他重要信息。对于 Argoverse 数据集,配置文件Argoverse.yaml
可以在以下链接找到:Argoverse.yaml。
欲了解有关 YAML 配置的更多信息,请参阅我们的数据集指南。
COCO 数据集
COCO(上下文中的常见对象)数据集是一个大规模对象检测、分割和字幕数据集。它旨在鼓励研究各种对象类别,并且通常用于计算机视觉模型的基准测试。对于从事对象检测、分割和姿态估计任务的研究人员和开发人员来说,它是一个必不可少的数据集。
www.youtube.com/embed/uDrn9QZJ2lk
Watch: Ultralytics COCO 数据集概述
COCO 预训练模型
| 模型 | 尺寸 ^((像素)) | mAP^(val 50-95) | 速度 ^(CPU ONNX
(ms)) | 速度 ^(A100 TensorRT
(ms)) | params ^((M)) | FLOPs ^((B)) |
--- | --- | --- | --- | --- | --- | --- |
---|---|---|---|---|---|---|
YOLOv8n | 640 | 37.3 | 80.4 | 0.99 | 3.2 | 8.7 |
YOLOv8s | 640 | 44.9 | 128.4 | 1.20 | 11.2 | 28.6 |
YOLOv8m | 640 | 50.2 | 234.7 | 1.83 | 25.9 | 78.9 |
YOLOv8l | 640 | 52.9 | 375.2 | 2.39 | 43.7 | 165.2 |
YOLOv8x | 640 | 53.9 | 479.1 | 3.53 | 68.2 | 257.8 |
主要特点
-
COCO 包含 330K 张图像,其中 200K 张图像具有对象检测、分割和字幕任务的注释。
-
数据集包括 80 个对象类别,包括常见对象如汽车、自行车和动物,以及更具体的类别,如雨伞、手提包和运动设备。
-
注释包括每个图像的对象边界框、分割蒙版和字幕。
-
COCO 提供了标准化的评估指标,如对象检测的平均精度(mAP)和分割任务的平均召回率(mAR),适合于比较模型性能。
数据集结构
COCO 数据集分为三个子集:
-
Train2017: 这个子集包含 118K 张用于训练对象检测、分割和字幕模型的图像。
-
Val2017: 这个子集包含用于模型训练验证目的的 5K 张图像。
-
Test2017: 这个子集包含用于测试和基准测试训练模型的 20K 张图像。该子集的地面实况标注并未公开,结果将提交至COCO 评估服务器进行性能评估。
应用
COCO 数据集广泛用于训练和评估深度学习模型,包括目标检测(如 YOLO、Faster R-CNN 和 SSD)、实例分割(如 Mask R-CNN)和关键点检测(如 OpenPose)。该数据集具有多样的对象类别集合、大量注释图像以及标准化的评估指标,使其成为计算机视觉研究人员和从业者的重要资源。
数据集 YAML
YAML(Yet Another Markup Language)文件用于定义数据集配置。它包含有关数据集路径、类别和其他相关信息的信息。在 COCO 数据集的情况下,coco.yaml
文件维护在 github.com/ultralytics/ultralytics/blob/main/ultralytics/cfg/datasets/coco.yaml
。
ultralytics/cfg/datasets/coco.yaml
# Ultralytics YOLO 🚀, AGPL-3.0 license
# COCO 2017 dataset https://cocodataset.org by Microsoft
# Documentation: https://docs.ultralytics.com/datasets/detect/coco/
# Example usage: yolo train data=coco.yaml
# parent
# ├── ultralytics
# └── datasets
# └── coco ← downloads here (20.1 GB)
# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
path: ../datasets/coco # dataset root dir
train: train2017.txt # train images (relative to 'path') 118287 images
val: val2017.txt # val images (relative to 'path') 5000 images
test: test-dev2017.txt # 20288 of 40670 images, submit to https://competitions.codalab.org/competitions/20794
# Classes
names:
0: person
1: bicycle
2: car
3: motorcycle
4: airplane
5: bus
6: train
7: truck
8: boat
9: traffic light
10: fire hydrant
11: stop sign
12: parking meter
13: bench
14: bird
15: cat
16: dog
17: horse
18: sheep
19: cow
20: elephant
21: bear
22: zebra
23: giraffe
24: backpack
25: umbrella
26: handbag
27: tie
28: suitcase
29: frisbee
30: skis
31: snowboard
32: sports ball
33: kite
34: baseball bat
35: baseball glove
36: skateboard
37: surfboard
38: tennis racket
39: bottle
40: wine glass
41: cup
42: fork
43: knife
44: spoon
45: bowl
46: banana
47: apple
48: sandwich
49: orange
50: broccoli
51: carrot
52: hot dog
53: pizza
54: donut
55: cake
56: chair
57: couch
58: potted plant
59: bed
60: dining table
61: toilet
62: tv
63: laptop
64: mouse
65: remote
66: keyboard
67: cell phone
68: microwave
69: oven
70: toaster
71: sink
72: refrigerator
73: book
74: clock
75: vase
76: scissors
77: teddy bear
78: hair drier
79: toothbrush
# Download script/URL (optional)
download: |
from ultralytics.utils.downloads import download
from pathlib import Path
# Download labels
segments = True # segment or box labels
dir = Path(yaml['path']) # dataset root dir
url = 'https://github.com/ultralytics/assets/releases/download/v0.0.0/'
urls = [url + ('coco2017labels-segments.zip' if segments else 'coco2017labels.zip')] # labels
download(urls, dir=dir.parent)
# Download data
urls = ['http://images.cocodataset.org/zips/train2017.zip', # 19G, 118k images
'http://images.cocodataset.org/zips/val2017.zip', # 1G, 5k images
'http://images.cocodataset.org/zips/test2017.zip'] # 7G, 41k images (optional)
download(urls, dir=dir / 'images', threads=3)
使用
要在 COCO 数据集上训练 100 个 epochs 的 YOLOv8n 模型,并使用 640 的图像大小,可以使用以下代码片段。有关可用参数的详细列表,请参阅模型训练页面。
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="coco.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=coco.yaml model=yolov8n.pt epochs=100 imgsz=640
样本图像和注释
COCO 数据集包含多样的图像集,具有各种对象类别和复杂场景。以下是数据集中的一些图像示例,以及它们的相应注释:
- 镶嵌图像:这幅图像展示了由镶嵌数据集图像组成的训练批次。镶嵌是训练过程中使用的一种技术,将多个图像合并成单个图像,以增加每个训练批次中对象和场景的多样性。这有助于提高模型对不同对象大小、长宽比和上下文的泛化能力。
该示例展示了 COCO 数据集中图像的多样性和复杂性,以及在训练过程中使用镶嵌技术的好处。
引用和致谢
如果您在研究或开发工作中使用 COCO 数据集,请引用以下论文:
@misc{lin2015microsoft,
title={Microsoft COCO: Common Objects in Context},
author={Tsung-Yi Lin and Michael Maire and Serge Belongie and Lubomir Bourdev and Ross Girshick and James Hays and Pietro Perona and Deva Ramanan and C. Lawrence Zitnick and Piotr Dollár},
year={2015},
eprint={1405.0312},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
我们希望感谢 COCO 联合体为计算机视觉社区创建和维护这一宝贵资源。有关 COCO 数据集及其创建者的更多信息,请访问COCO 数据集网站。
常见问题
COCO 数据集是什么,对计算机视觉的重要性在哪里?
COCO 数据集(上下文中的常见对象)是用于目标检测、分割和字幕的大规模数据集。它包含了 33 万张图像,并对 80 种对象类别进行了详细的注释,因此对于基准测试和训练计算机视觉模型至关重要。研究人员使用 COCO 数据集,因为它包含多样的类别和标准化的评估指标,如平均精度(mAP)。
如何使用 COCO 数据集训练 YOLO 模型?
要使用 COCO 数据集训练 YOLOv8 模型,可以使用以下代码片段:
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="coco.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=coco.yaml model=yolov8n.pt epochs=100 imgsz=640
参考训练页面以获取更多关于可用参数的详细信息。
COCO 数据集的关键特征是什么?
COCO 数据集包括:
-
包括 330K 张图像,其中有 200K 张用于目标检测、分割和字幕。
-
包括 80 个物体类别,从常见物品如汽车和动物到特定物品如手提包和运动装备。
-
标准化的目标检测评估指标(mAP)和分割评估指标(平均召回率 mAR)。
-
Mosaicing 技术用于训练批次,以增强模型对各种物体尺寸和背景的泛化能力。
在哪里可以找到在 COCO 数据集上训练的预训练 YOLOv8 模型?
在文档中提供的链接可以下载在 COCO 数据集上预训练的 YOLOv8 模型。例如:
这些模型在大小、mAP 和推理速度上各有不同,为不同性能和资源需求提供了选择。
COCO 数据集的结构及其使用方法?
COCO 数据集分为三个子集:
-
Train2017: 用于训练的 118K 张图像。
-
Val2017: 用于训练验证的 5K 张图像。
-
Test2017: 用于评估训练模型的 20K 张图像。需将结果提交至COCO 评估服务器进行性能评估。
数据集的 YAML 配置文件可在coco.yaml找到,定义了路径、类别和数据集的详细信息。
LVIS 数据集
LVIS 数据集 是由 Facebook AI Research(FAIR)开发和发布的大规模、细粒度词汇级别注释数据集,主要用作物体检测和实例分割的研究基准,具有大量类别的词汇,旨在推动计算机视觉领域的进一步发展。
www.youtube.com/embed/cfTKj96TjSE
Watch: YOLO World 使用 LVIS 数据集的训练工作流程
主要特点
-
LVIS 包含 160k 张图像和 2M 个实例标注,用于物体检测、分割和字幕任务。
-
该数据集包括 1203 个对象类别,包括常见对象如汽车、自行车和动物,以及更具体的类别如雨伞、手提包和体育设备。
-
标注包括每张图像的对象边界框、分割蒙版和说明。
-
LVIS 提供标准化的评估指标,如物体检测的平均精确度(mAP)和分割任务的平均召回率(mAR),适合比较模型性能。
-
LVIS 使用与 COCO 数据集完全相同的图像,但具有不同的拆分和不同的注释。
数据集结构
LVIS 数据集分为三个子集:
-
Train: 这个子集包含 100k 张图像,用于训练物体检测、分割和字幕模型。
-
Val: 这个子集有 20k 张图像,用于模型训练的验证目的。
-
Minival: 这个子集与 COCO val2017 集合完全相同,有 5k 张图像,用于模型训练的验证目的。
-
Test: 这个子集包含 20k 张图像,用于测试和基准测试经过训练的模型。此子集的地面真实标注不公开,结果提交到 LVIS 评估服务器 进行性能评估。
应用
LVIS 数据集被广泛用于训练和评估物体检测(如 YOLO、Faster R-CNN 和 SSD)、实例分割(如 Mask R-CNN)的深度学习模型。数据集的多样的对象类别集合、大量注释图像和标准化评估指标使其成为计算机视觉研究人员和从业者的重要资源。
数据集 YAML
使用 YAML(另一种标记语言)文件定义数据集配置。它包含关于数据集路径、类别和其他相关信息的信息。在 LVIS 数据集的情况下,lvis.yaml
文件保存在 github.com/ultralytics/ultralytics/blob/main/ultralytics/cfg/datasets/lvis.yaml
。
ultralytics/cfg/datasets/lvis.yaml
# Ultralytics YOLO 🚀, AGPL-3.0 license
# LVIS dataset http://www.lvisdataset.org by Facebook AI Research.
# Documentation: https://docs.ultralytics.com/datasets/detect/lvis/
# Example usage: yolo train data=lvis.yaml
# parent
# ├── ultralytics
# └── datasets
# └── lvis ← downloads here (20.1 GB)
# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
path: ../datasets/lvis # dataset root dir
train: train.txt # train images (relative to 'path') 100170 images
val: val.txt # val images (relative to 'path') 19809 images
minival: minival.txt # minval images (relative to 'path') 5000 images
names:
0: aerosol can/spray can
1: air conditioner
2: airplane/aeroplane
3: alarm clock
4: alcohol/alcoholic beverage
5: alligator/gator
6: almond
7: ambulance
8: amplifier
9: anklet/ankle bracelet
10: antenna/aerial/transmitting aerial
11: apple
12: applesauce
13: apricot
14: apron
15: aquarium/fish tank
16: arctic/arctic type of shoe/galosh/golosh/rubber/rubber type of shoe/gumshoe
17: armband
18: armchair
19: armoire
20: armor/armour
21: artichoke
22: trash can/garbage can/wastebin/dustbin/trash barrel/trash bin
23: ashtray
24: asparagus
25: atomizer/atomiser/spray/sprayer/nebulizer/nebuliser
26: avocado
27: award/accolade
28: awning
29: ax/axe
30: baboon
31: baby buggy/baby carriage/perambulator/pram/stroller
32: basketball backboard
33: backpack/knapsack/packsack/rucksack/haversack
34: handbag/purse/pocketbook
35: suitcase/baggage/luggage
36: bagel/beigel
37: bagpipe
38: baguet/baguette
39: bait/lure
40: ball
41: ballet skirt/tutu
42: balloon
43: bamboo
44: banana
45: Band Aid
46: bandage
47: bandanna/bandana
48: banjo
49: banner/streamer
50: barbell
51: barge
52: barrel/cask
53: barrette
54: barrow/garden cart/lawn cart/wheelbarrow
55: baseball base
56: baseball
57: baseball bat
58: baseball cap/jockey cap/golf cap
59: baseball glove/baseball mitt
60: basket/handbasket
61: basketball
62: bass horn/sousaphone/tuba
63: bat/bat animal
64: bath mat
65: bath towel
66: bathrobe
67: bathtub/bathing tub
68: batter/batter food
69: battery
70: beachball
71: bead
72: bean curd/tofu
73: beanbag
74: beanie/beany
75: bear
76: bed
77: bedpan
78: bedspread/bedcover/bed covering/counterpane/spread
79: cow
80: beef/beef food/boeuf/boeuf food
81: beeper/pager
82: beer bottle
83: beer can
84: beetle
85: bell
86: bell pepper/capsicum
87: belt
88: belt buckle
89: bench
90: beret
91: bib
92: Bible
93: bicycle/bike/bike bicycle
94: visor/vizor
95: billboard
96: binder/ring-binder
97: binoculars/field glasses/opera glasses
98: bird
99: birdfeeder
100: birdbath
101: birdcage
102: birdhouse
103: birthday cake
104: birthday card
105: pirate flag
106: black sheep
107: blackberry
108: blackboard/chalkboard
109: blanket
110: blazer/sport jacket/sport coat/sports jacket/sports coat
111: blender/liquidizer/liquidiser
112: blimp
113: blinker/flasher
114: blouse
115: blueberry
116: gameboard
117: boat/ship/ship boat
118: bob/bobber/bobfloat
119: bobbin/spool/reel
120: bobby pin/hairgrip
121: boiled egg/coddled egg
122: bolo tie/bolo/bola tie/bola
123: deadbolt
124: bolt
125: bonnet
126: book
127: bookcase
128: booklet/brochure/leaflet/pamphlet
129: bookmark/bookmarker
130: boom microphone/microphone boom
131: boot
132: bottle
133: bottle opener
134: bouquet
135: bow/bow weapon
136: bow/bow decorative ribbons
137: bow-tie/bowtie
138: bowl
139: pipe bowl
140: bowler hat/bowler/derby hat/derby/plug hat
141: bowling ball
142: box
143: boxing glove
144: suspenders
145: bracelet/bangle
146: brass plaque
147: brassiere/bra/bandeau
148: bread-bin/breadbox
149: bread
150: breechcloth/breechclout/loincloth
151: bridal gown/wedding gown/wedding dress
152: briefcase
153: broccoli
154: broach
155: broom
156: brownie
157: brussels sprouts
158: bubble gum
159: bucket/pail
160: horse buggy
161: horned cow
162: bulldog
163: bulldozer/dozer
164: bullet train
165: bulletin board/notice board
166: bulletproof vest
167: bullhorn/megaphone
168: bun/roll
169: bunk bed
170: buoy
171: burrito
172: bus/bus vehicle/autobus/charabanc/double-decker/motorbus/motorcoach
173: business card
174: butter
175: butterfly
176: button
177: cab/cab taxi/taxi/taxicab
178: cabana
179: cabin car/caboose
180: cabinet
181: locker/storage locker
182: cake
183: calculator
184: calendar
185: calf
186: camcorder
187: camel
188: camera
189: camera lens
190: camper/camper vehicle/camping bus/motor home
191: can/tin can
192: can opener/tin opener
193: candle/candlestick
194: candle holder
195: candy bar
196: candy cane
197: walking cane
198: canister/canister
199: canoe
200: cantaloup/cantaloupe
201: canteen
202: cap/cap headwear
203: bottle cap/cap/cap container lid
204: cape
205: cappuccino/coffee cappuccino
206: car/car automobile/auto/auto automobile/automobile
207: railcar/railcar part of a train/railway car/railway car part of a train/railroad car/railroad car part of a train
208: elevator car
209: car battery/automobile battery
210: identity card
211: card
212: cardigan
213: cargo ship/cargo vessel
214: carnation
215: horse carriage
216: carrot
217: tote bag
218: cart
219: carton
220: cash register/register/register for cash transactions
221: casserole
222: cassette
223: cast/plaster cast/plaster bandage
224: cat
225: cauliflower
226: cayenne/cayenne spice/cayenne pepper/cayenne pepper spice/red pepper/red pepper spice
227: CD player
228: celery
229: cellular telephone/cellular phone/cellphone/mobile phone/smart phone
230: chain mail/ring mail/chain armor/chain armour/ring armor/ring armour
231: chair
232: chaise longue/chaise/daybed
233: chalice
234: chandelier
235: chap
236: checkbook/chequebook
237: checkerboard
238: cherry
239: chessboard
240: chicken/chicken animal
241: chickpea/garbanzo
242: chili/chili vegetable/chili pepper/chili pepper vegetable/chilli/chilli vegetable/chilly/chilly vegetable/chile/chile vegetable
243: chime/gong
244: chinaware
245: crisp/crisp potato chip/potato chip
246: poker chip
247: chocolate bar
248: chocolate cake
249: chocolate milk
250: chocolate mousse
251: choker/collar/neckband
252: chopping board/cutting board/chopping block
253: chopstick
254: Christmas tree
255: slide
256: cider/cyder
257: cigar box
258: cigarette
259: cigarette case/cigarette pack
260: cistern/water tank
261: clarinet
262: clasp
263: cleansing agent/cleanser/cleaner
264: cleat/cleat for securing rope
265: clementine
266: clip
267: clipboard
268: clippers/clippers for plants
269: cloak
270: clock/timepiece/timekeeper
271: clock tower
272: clothes hamper/laundry basket/clothes basket
273: clothespin/clothes peg
274: clutch bag
275: coaster
276: coat
277: coat hanger/clothes hanger/dress hanger
278: coatrack/hatrack
279: cock/rooster
280: cockroach
281: cocoa/cocoa beverage/hot chocolate/hot chocolate beverage/drinking chocolate
282: coconut/cocoanut
283: coffee maker/coffee machine
284: coffee table/cocktail table
285: coffeepot
286: coil
287: coin
288: colander/cullender
289: coleslaw/slaw
290: coloring material/colouring material
291: combination lock
292: pacifier/teething ring
293: comic book
294: compass
295: computer keyboard/keyboard/keyboard computer
296: condiment
297: cone/traffic cone
298: control/controller
299: convertible/convertible automobile
300: sofa bed
301: cooker
302: cookie/cooky/biscuit/biscuit cookie
303: cooking utensil
304: cooler/cooler for food/ice chest
305: cork/cork bottle plug/bottle cork
306: corkboard
307: corkscrew/bottle screw
308: edible corn/corn/maize
309: cornbread
310: cornet/horn/trumpet
311: cornice/valance/valance board/pelmet
312: cornmeal
313: corset/girdle
314: costume
315: cougar/puma/catamount/mountain lion/panther
316: coverall
317: cowbell
318: cowboy hat/ten-gallon hat
319: crab/crab animal
320: crabmeat
321: cracker
322: crape/crepe/French pancake
323: crate
324: crayon/wax crayon
325: cream pitcher
326: crescent roll/croissant
327: crib/cot
328: crock pot/earthenware jar
329: crossbar
330: crouton
331: crow
332: crowbar/wrecking bar/pry bar
333: crown
334: crucifix
335: cruise ship/cruise liner
336: police cruiser/patrol car/police car/squad car
337: crumb
338: crutch
339: cub/cub animal
340: cube/square block
341: cucumber/cuke
342: cufflink
343: cup
344: trophy cup
345: cupboard/closet
346: cupcake
347: hair curler/hair roller/hair crimper
348: curling iron
349: curtain/drapery
350: cushion
351: cylinder
352: cymbal
353: dagger
354: dalmatian
355: dartboard
356: date/date fruit
357: deck chair/beach chair
358: deer/cervid
359: dental floss/floss
360: desk
361: detergent
362: diaper
363: diary/journal
364: die/dice
365: dinghy/dory/rowboat
366: dining table
367: tux/tuxedo
368: dish
369: dish antenna
370: dishrag/dishcloth
371: dishtowel/tea towel
372: dishwasher/dishwashing machine
373: dishwasher detergent/dishwashing detergent/dishwashing liquid/dishsoap
374: dispenser
375: diving board
376: Dixie cup/paper cup
377: dog
378: dog collar
379: doll
380: dollar/dollar bill/one dollar bill
381: dollhouse/doll's house
382: dolphin
383: domestic ass/donkey
384: doorknob/doorhandle
385: doormat/welcome mat
386: doughnut/donut
387: dove
388: dragonfly
389: drawer
390: underdrawers/boxers/boxershorts
391: dress/frock
392: dress hat/high hat/opera hat/silk hat/top hat
393: dress suit
394: dresser
395: drill
396: drone
397: dropper/eye dropper
398: drum/drum musical instrument
399: drumstick
400: duck
401: duckling
402: duct tape
403: duffel bag/duffle bag/duffel/duffle
404: dumbbell
405: dumpster
406: dustpan
407: eagle
408: earphone/earpiece/headphone
409: earplug
410: earring
411: easel
412: eclair
413: eel
414: egg/eggs
415: egg roll/spring roll
416: egg yolk/yolk/yolk egg
417: eggbeater/eggwhisk
418: eggplant/aubergine
419: electric chair
420: refrigerator
421: elephant
422: elk/moose
423: envelope
424: eraser
425: escargot
426: eyepatch
427: falcon
428: fan
429: faucet/spigot/tap
430: fedora
431: ferret
432: Ferris wheel
433: ferry/ferryboat
434: fig/fig fruit
435: fighter jet/fighter aircraft/attack aircraft
436: figurine
437: file cabinet/filing cabinet
438: file/file tool
439: fire alarm/smoke alarm
440: fire engine/fire truck
441: fire extinguisher/extinguisher
442: fire hose
443: fireplace
444: fireplug/fire hydrant/hydrant
445: first-aid kit
446: fish
447: fish/fish food
448: fishbowl/goldfish bowl
449: fishing rod/fishing pole
450: flag
451: flagpole/flagstaff
452: flamingo
453: flannel
454: flap
455: flash/flashbulb
456: flashlight/torch
457: fleece
458: flip-flop/flip-flop sandal
459: flipper/flipper footwear/fin/fin footwear
460: flower arrangement/floral arrangement
461: flute glass/champagne flute
462: foal
463: folding chair
464: food processor
465: football/football American
466: football helmet
467: footstool/footrest
468: fork
469: forklift
470: freight car
471: French toast
472: freshener/air freshener
473: frisbee
474: frog/toad/toad frog
475: fruit juice
476: frying pan/frypan/skillet
477: fudge
478: funnel
479: futon
480: gag/muzzle
481: garbage
482: garbage truck
483: garden hose
484: gargle/mouthwash
485: gargoyle
486: garlic/ail
487: gasmask/respirator/gas helmet
488: gazelle
489: gelatin/jelly
490: gemstone
491: generator
492: giant panda/panda/panda bear
493: gift wrap
494: ginger/gingerroot
495: giraffe
496: cincture/sash/waistband/waistcloth
497: glass/glass drink container/drinking glass
498: globe
499: glove
500: goat
501: goggles
502: goldfish
503: golf club/golf-club
504: golfcart
505: gondola/gondola boat
506: goose
507: gorilla
508: gourd
509: grape
510: grater
511: gravestone/headstone/tombstone
512: gravy boat/gravy holder
513: green bean
514: green onion/spring onion/scallion
515: griddle
516: grill/grille/grillwork/radiator grille
517: grits/hominy grits
518: grizzly/grizzly bear
519: grocery bag
520: guitar
521: gull/seagull
522: gun
523: hairbrush
524: hairnet
525: hairpin
526: halter top
527: ham/jambon/gammon
528: hamburger/beefburger/burger
529: hammer
530: hammock
531: hamper
532: hamster
533: hair dryer
534: hand glass/hand mirror
535: hand towel/face towel
536: handcart/pushcart/hand truck
537: handcuff
538: handkerchief
539: handle/grip/handgrip
540: handsaw/carpenter's saw
541: hardback book/hardcover book
542: harmonium/organ/organ musical instrument/reed organ/reed organ musical instrument
543: hat
544: hatbox
545: veil
546: headband
547: headboard
548: headlight/headlamp
549: headscarf
550: headset
551: headstall/headstall for horses/headpiece/headpiece for horses
552: heart
553: heater/warmer
554: helicopter
555: helmet
556: heron
557: highchair/feeding chair
558: hinge
559: hippopotamus
560: hockey stick
561: hog/pig
562: home plate/home plate baseball/home base/home base baseball
563: honey
564: fume hood/exhaust hood
565: hook
566: hookah/narghile/nargileh/sheesha/shisha/water pipe
567: hornet
568: horse
569: hose/hosepipe
570: hot-air balloon
571: hotplate
572: hot sauce
573: hourglass
574: houseboat
575: hummingbird
576: hummus/humus/hommos/hoummos/humous
577: polar bear
578: icecream
579: popsicle
580: ice maker
581: ice pack/ice bag
582: ice skate
583: igniter/ignitor/lighter
584: inhaler/inhalator
585: iPod
586: iron/iron for clothing/smoothing iron/smoothing iron for clothing
587: ironing board
588: jacket
589: jam
590: jar
591: jean/blue jean/denim
592: jeep/landrover
593: jelly bean/jelly egg
594: jersey/T-shirt/tee shirt
595: jet plane/jet-propelled plane
596: jewel/gem/precious stone
597: jewelry/jewellery
598: joystick
599: jumpsuit
600: kayak
601: keg
602: kennel/doghouse
603: kettle/boiler
604: key
605: keycard
606: kilt
607: kimono
608: kitchen sink
609: kitchen table
610: kite
611: kitten/kitty
612: kiwi fruit
613: knee pad
614: knife
615: knitting needle
616: knob
617: knocker/knocker on a door/doorknocker
618: koala/koala bear
619: lab coat/laboratory coat
620: ladder
621: ladle
622: ladybug/ladybeetle/ladybird beetle
623: lamb/lamb animal
624: lamb-chop/lambchop
625: lamp
626: lamppost
627: lampshade
628: lantern
629: lanyard/laniard
630: laptop computer/notebook computer
631: lasagna/lasagne
632: latch
633: lawn mower
634: leather
635: legging/legging clothing/leging/leging clothing/leg covering
636: Lego/Lego set
637: legume
638: lemon
639: lemonade
640: lettuce
641: license plate/numberplate
642: life buoy/lifesaver/life belt/life ring
643: life jacket/life vest
644: lightbulb
645: lightning rod/lightning conductor
646: lime
647: limousine
648: lion
649: lip balm
650: liquor/spirits/hard liquor/liqueur/cordial
651: lizard
652: log
653: lollipop
654: speaker/speaker stereo equipment
655: loveseat
656: machine gun
657: magazine
658: magnet
659: mail slot
660: mailbox/mailbox at home/letter box/letter box at home
661: mallard
662: mallet
663: mammoth
664: manatee
665: mandarin orange
666: manager/through
667: manhole
668: map
669: marker
670: martini
671: mascot
672: mashed potato
673: masher
674: mask/facemask
675: mast
676: mat/mat gym equipment/gym mat
677: matchbox
678: mattress
679: measuring cup
680: measuring stick/ruler/ruler measuring stick/measuring rod
681: meatball
682: medicine
683: melon
684: microphone
685: microscope
686: microwave oven
687: milestone/milepost
688: milk
689: milk can
690: milkshake
691: minivan
692: mint candy
693: mirror
694: mitten
695: mixer/mixer kitchen tool/stand mixer
696: money
697: monitor/monitor computer equipment
698: monkey
699: motor
700: motor scooter/scooter
701: motor vehicle/automotive vehicle
702: motorcycle
703: mound/mound baseball/pitcher's mound
704: mouse/mouse computer equipment/computer mouse
705: mousepad
706: muffin
707: mug
708: mushroom
709: music stool/piano stool
710: musical instrument/instrument/instrument musical
711: nailfile
712: napkin/table napkin/serviette
713: neckerchief
714: necklace
715: necktie/tie/tie necktie
716: needle
717: nest
718: newspaper/paper/paper newspaper
719: newsstand
720: nightshirt/nightwear/sleepwear/nightclothes
721: nosebag/nosebag for animals/feedbag
722: noseband/noseband for animals/nosepiece/nosepiece for animals
723: notebook
724: notepad
725: nut
726: nutcracker
727: oar
728: octopus/octopus food
729: octopus/octopus animal
730: oil lamp/kerosene lamp/kerosine lamp
731: olive oil
732: omelet/omelette
733: onion
734: orange/orange fruit
735: orange juice
736: ostrich
737: ottoman/pouf/pouffe/hassock
738: oven
739: overalls/overalls clothing
740: owl
741: packet
742: inkpad/inking pad/stamp pad
743: pad
744: paddle/boat paddle
745: padlock
746: paintbrush
747: painting
748: pajamas/pyjamas
749: palette/pallet
750: pan/pan for cooking/cooking pan
751: pan/pan metal container
752: pancake
753: pantyhose
754: papaya
755: paper plate
756: paper towel
757: paperback book/paper-back book/softback book/soft-cover book
758: paperweight
759: parachute
760: parakeet/parrakeet/parroket/paraquet/paroquet/parroquet
761: parasail/parasail sports
762: parasol/sunshade
763: parchment
764: parka/anorak
765: parking meter
766: parrot
767: passenger car/passenger car part of a train/coach/coach part of a train
768: passenger ship
769: passport
770: pastry
771: patty/patty food
772: pea/pea food
773: peach
774: peanut butter
775: pear
776: peeler/peeler tool for fruit and vegetables
777: wooden leg/pegleg
778: pegboard
779: pelican
780: pen
781: pencil
782: pencil box/pencil case
783: pencil sharpener
784: pendulum
785: penguin
786: pennant
787: penny/penny coin
788: pepper/peppercorn
789: pepper mill/pepper grinder
790: perfume
791: persimmon
792: person/baby/child/boy/girl/man/woman/human
793: pet
794: pew/pew church bench/church bench
795: phonebook/telephone book/telephone directory
796: phonograph record/phonograph recording/record/record phonograph recording
797: piano
798: pickle
799: pickup truck
800: pie
801: pigeon
802: piggy bank/penny bank
803: pillow
804: pin/pin non jewelry
805: pineapple
806: pinecone
807: ping-pong ball
808: pinwheel
809: tobacco pipe
810: pipe/piping
811: pistol/handgun
812: pita/pita bread/pocket bread
813: pitcher/pitcher vessel for liquid/ewer
814: pitchfork
815: pizza
816: place mat
817: plate
818: platter
819: playpen
820: pliers/plyers
821: plow/plow farm equipment/plough/plough farm equipment
822: plume
823: pocket watch
824: pocketknife
825: poker/poker fire stirring tool/stove poker/fire hook
826: pole/post
827: polo shirt/sport shirt
828: poncho
829: pony
830: pool table/billiard table/snooker table
831: pop/pop soda/soda/soda pop/tonic/soft drink
832: postbox/postbox public/mailbox/mailbox public
833: postcard/postal card/mailing-card
834: poster/placard
835: pot
836: flowerpot
837: potato
838: potholder
839: pottery/clayware
840: pouch
841: power shovel/excavator/digger
842: prawn/shrimp
843: pretzel
844: printer/printing machine
845: projectile/projectile weapon/missile
846: projector
847: propeller/propellor
848: prune
849: pudding
850: puffer/puffer fish/pufferfish/blowfish/globefish
851: puffin
852: pug-dog
853: pumpkin
854: puncher
855: puppet/marionette
856: puppy
857: quesadilla
858: quiche
859: quilt/comforter
860: rabbit
861: race car/racing car
862: racket/racquet
863: radar
864: radiator
865: radio receiver/radio set/radio/tuner/tuner radio
866: radish/daikon
867: raft
868: rag doll
869: raincoat/waterproof jacket
870: ram/ram animal
871: raspberry
872: rat
873: razorblade
874: reamer/reamer juicer/juicer/juice reamer
875: rearview mirror
876: receipt
877: recliner/reclining chair/lounger/lounger chair
878: record player/phonograph/phonograph record player/turntable
879: reflector
880: remote control
881: rhinoceros
882: rib/rib food
883: rifle
884: ring
885: river boat
886: road map
887: robe
888: rocking chair
889: rodent
890: roller skate
891: Rollerblade
892: rolling pin
893: root beer
894: router/router computer equipment
895: rubber band/elastic band
896: runner/runner carpet
897: plastic bag/paper bag
898: saddle/saddle on an animal
899: saddle blanket/saddlecloth/horse blanket
900: saddlebag
901: safety pin
902: sail
903: salad
904: salad plate/salad bowl
905: salami
906: salmon/salmon fish
907: salmon/salmon food
908: salsa
909: saltshaker
910: sandal/sandal type of shoe
911: sandwich
912: satchel
913: saucepan
914: saucer
915: sausage
916: sawhorse/sawbuck
917: saxophone
918: scale/scale measuring instrument
919: scarecrow/strawman
920: scarf
921: school bus
922: scissors
923: scoreboard
924: scraper
925: screwdriver
926: scrubbing brush
927: sculpture
928: seabird/seafowl
929: seahorse
930: seaplane/hydroplane
931: seashell
932: sewing machine
933: shaker
934: shampoo
935: shark
936: sharpener
937: Sharpie
938: shaver/shaver electric/electric shaver/electric razor
939: shaving cream/shaving soap
940: shawl
941: shears
942: sheep
943: shepherd dog/sheepdog
944: sherbert/sherbet
945: shield
946: shirt
947: shoe/sneaker/sneaker type of shoe/tennis shoe
948: shopping bag
949: shopping cart
950: short pants/shorts/shorts clothing/trunks/trunks clothing
951: shot glass
952: shoulder bag
953: shovel
954: shower head
955: shower cap
956: shower curtain
957: shredder/shredder for paper
958: signboard
959: silo
960: sink
961: skateboard
962: skewer
963: ski
964: ski boot
965: ski parka/ski jacket
966: ski pole
967: skirt
968: skullcap
969: sled/sledge/sleigh
970: sleeping bag
971: sling/sling bandage/triangular bandage
972: slipper/slipper footwear/carpet slipper/carpet slipper footwear
973: smoothie
974: snake/serpent
975: snowboard
976: snowman
977: snowmobile
978: soap
979: soccer ball
980: sock
981: sofa/couch/lounge
982: softball
983: solar array/solar battery/solar panel
984: sombrero
985: soup
986: soup bowl
987: soupspoon
988: sour cream/soured cream
989: soya milk/soybean milk/soymilk
990: space shuttle
991: sparkler/sparkler fireworks
992: spatula
993: spear/lance
994: spectacles/specs/eyeglasses/glasses
995: spice rack
996: spider
997: crawfish/crayfish
998: sponge
999: spoon
1000: sportswear/athletic wear/activewear
1001: spotlight
1002: squid/squid food/calamari/calamary
1003: squirrel
1004: stagecoach
1005: stapler/stapler stapling machine
1006: starfish/sea star
1007: statue/statue sculpture
1008: steak/steak food
1009: steak knife
1010: steering wheel
1011: stepladder
1012: step stool
1013: stereo/stereo sound system
1014: stew
1015: stirrer
1016: stirrup
1017: stool
1018: stop sign
1019: brake light
1020: stove/kitchen stove/range/range kitchen appliance/kitchen range/cooking stove
1021: strainer
1022: strap
1023: straw/straw for drinking/drinking straw
1024: strawberry
1025: street sign
1026: streetlight/street lamp
1027: string cheese
1028: stylus
1029: subwoofer
1030: sugar bowl
1031: sugarcane/sugarcane plant
1032: suit/suit clothing
1033: sunflower
1034: sunglasses
1035: sunhat
1036: surfboard
1037: sushi
1038: mop
1039: sweat pants
1040: sweatband
1041: sweater
1042: sweatshirt
1043: sweet potato
1044: swimsuit/swimwear/bathing suit/swimming costume/bathing costume/swimming trunks/bathing trunks
1045: sword
1046: syringe
1047: Tabasco sauce
1048: table-tennis table/ping-pong table
1049: table
1050: table lamp
1051: tablecloth
1052: tachometer
1053: taco
1054: tag
1055: taillight/rear light
1056: tambourine
1057: army tank/armored combat vehicle/armoured combat vehicle
1058: tank/tank storage vessel/storage tank
1059: tank top/tank top clothing
1060: tape/tape sticky cloth or paper
1061: tape measure/measuring tape
1062: tapestry
1063: tarp
1064: tartan/plaid
1065: tassel
1066: tea bag
1067: teacup
1068: teakettle
1069: teapot
1070: teddy bear
1071: telephone/phone/telephone set
1072: telephone booth/phone booth/call box/telephone box/telephone kiosk
1073: telephone pole/telegraph pole/telegraph post
1074: telephoto lens/zoom lens
1075: television camera/tv camera
1076: television set/tv/tv set
1077: tennis ball
1078: tennis racket
1079: tequila
1080: thermometer
1081: thermos bottle
1082: thermostat
1083: thimble
1084: thread/yarn
1085: thumbtack/drawing pin/pushpin
1086: tiara
1087: tiger
1088: tights/tights clothing/leotards
1089: timer/stopwatch
1090: tinfoil
1091: tinsel
1092: tissue paper
1093: toast/toast food
1094: toaster
1095: toaster oven
1096: toilet
1097: toilet tissue/toilet paper/bathroom tissue
1098: tomato
1099: tongs
1100: toolbox
1101: toothbrush
1102: toothpaste
1103: toothpick
1104: cover
1105: tortilla
1106: tow truck
1107: towel
1108: towel rack/towel rail/towel bar
1109: toy
1110: tractor/tractor farm equipment
1111: traffic light
1112: dirt bike
1113: trailer truck/tractor trailer/trucking rig/articulated lorry/semi truck
1114: train/train railroad vehicle/railroad train
1115: trampoline
1116: tray
1117: trench coat
1118: triangle/triangle musical instrument
1119: tricycle
1120: tripod
1121: trousers/pants/pants clothing
1122: truck
1123: truffle/truffle chocolate/chocolate truffle
1124: trunk
1125: vat
1126: turban
1127: turkey/turkey food
1128: turnip
1129: turtle
1130: turtleneck/turtleneck clothing/polo-neck
1131: typewriter
1132: umbrella
1133: underwear/underclothes/underclothing/underpants
1134: unicycle
1135: urinal
1136: urn
1137: vacuum cleaner
1138: vase
1139: vending machine
1140: vent/blowhole/air vent
1141: vest/waistcoat
1142: videotape
1143: vinegar
1144: violin/fiddle
1145: vodka
1146: volleyball
1147: vulture
1148: waffle
1149: waffle iron
1150: wagon
1151: wagon wheel
1152: walking stick
1153: wall clock
1154: wall socket/wall plug/electric outlet/electrical outlet/outlet/electric receptacle
1155: wallet/billfold
1156: walrus
1157: wardrobe
1158: washbasin/basin/basin for washing/washbowl/washstand/handbasin
1159: automatic washer/washing machine
1160: watch/wristwatch
1161: water bottle
1162: water cooler
1163: water faucet/water tap/tap/tap water faucet
1164: water heater/hot-water heater
1165: water jug
1166: water gun/squirt gun
1167: water scooter/sea scooter/jet ski
1168: water ski
1169: water tower
1170: watering can
1171: watermelon
1172: weathervane/vane/vane weathervane/wind vane
1173: webcam
1174: wedding cake/bridecake
1175: wedding ring/wedding band
1176: wet suit
1177: wheel
1178: wheelchair
1179: whipped cream
1180: whistle
1181: wig
1182: wind chime
1183: windmill
1184: window box/window box for plants
1185: windshield wiper/windscreen wiper/wiper/wiper for windshield or screen
1186: windsock/air sock/air-sleeve/wind sleeve/wind cone
1187: wine bottle
1188: wine bucket/wine cooler
1189: wineglass
1190: blinder/blinder for horses
1191: wok
1192: wolf
1193: wooden spoon
1194: wreath
1195: wrench/spanner
1196: wristband
1197: wristlet/wrist band
1198: yacht
1199: yogurt/yoghurt/yoghourt
1200: yoke/yoke animal equipment
1201: zebra
1202: zucchini/courgette
# Download script/URL (optional)
download: |
from ultralytics.utils.downloads import download
from pathlib import Path
# Download labels
dir = Path(yaml['path']) # dataset root dir
url = 'https://github.com/ultralytics/assets/releases/download/v0.0.0/'
urls = [url + 'lvis-labels-segments.zip'] # labels
download(urls, dir=dir.parent)
# Download data
urls = ['http://images.cocodataset.org/zips/train2017.zip', # 19G, 118k images
'http://images.cocodataset.org/zips/val2017.zip', # 1G, 5k images
'http://images.cocodataset.org/zips/test2017.zip'] # 7G, 41k images (optional)
download(urls, dir=dir / 'images', threads=3)
使用方法
要在 LVIS 数据集上使用 640 像素大小的图像训练 100 个 epochs 的 YOLOv8n 模型,您可以使用以下代码片段。有关可用参数的详细列表,请参阅模型训练页面。
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="lvis.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=lvis.yaml model=yolov8n.pt epochs=100 imgsz=640
样本图像和注释
LVIS 数据集包含各种对象类别和复杂场景的多样化图像。以下是数据集中一些图像及其相应的注释示例:
- 马赛克图像:这幅图展示了由马赛克数据集图像组成的训练批次。马赛克是一种在训练过程中使用的技术,将多个图像合并成一张图像,以增加每个训练批次中对象和场景的多样性。这有助于改善模型对不同对象大小、长宽比和上下文的泛化能力。
此示例展示了 LVIS 数据集中图像的多样性和复杂性,以及在训练过程中使用马赛克的好处。
引用和致谢
如果您在研究或开发工作中使用 LVIS 数据集,请引用以下论文:
@inproceedings{gupta2019lvis,
title={{LVIS}: A Dataset for Large Vocabulary Instance Segmentation},
author={Gupta, Agrim and Dollar, Piotr and Girshick, Ross},
booktitle={Proceedings of the {IEEE} Conference on Computer Vision and Pattern Recognition},
year={2019}
}
我们要感谢 LVIS 联盟为计算机视觉社区创建和维护这一宝贵资源。有关 LVIS 数据集及其创建者的更多信息,请访问LVIS 数据集网站。
常见问题
LVIS 数据集是什么,如何在计算机视觉中使用?
LVIS 数据集是由 Facebook AI Research (FAIR)开发的带有细粒度词汇级注释的大规模数据集。它主要用于对象检测和实例分割,涵盖了超过 1203 个对象类别和 200 万个实例注释。研究人员和实践者使用它来训练和评估像 Ultralytics YOLO 这样的模型,用于高级计算机视觉任务。数据集的广泛大小和多样性使其成为推动检测和分割模型性能边界的重要资源。
如何使用 LVIS 数据集训练 YOLOv8n 模型?
要在 LVIS 数据集上使用 640 像素大小的图像训练 100 个 epochs 的 YOLOv8n 模型,请参考以下示例。此过程利用了 Ultralytics 的框架,提供了全面的训练功能。
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="lvis.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=lvis.yaml model=yolov8n.pt epochs=100 imgsz=640
如需详细的训练配置,请参阅训练文档。
LVIS 数据集与 COCO 数据集有何不同?
LVIS 数据集中的图像与 COCO 数据集中的图像相同,但两者在分割和注释方面有所不同。LVIS 提供了 1203 个对象类别的更大和更详细的词汇表,而 COCO 只有 80 个类别。此外,LVIS 侧重于注释的完整性和多样性,旨在通过提供更细致和全面的数据来推动对象检测和实例分割模型的极限。
为什么要在 LVIS 数据集上使用 Ultralytics YOLO 进行训练?
Ultralytics YOLO 模型,包括最新的 YOLOv8,针对实时目标检测进行了优化,具有领先的准确性和速度。它们支持广泛的注释,例如 LVIS 数据集提供的精细注释,使其成为高级计算机视觉应用的理想选择。此外,Ultralytics 提供与各种训练、验证和预测模式的无缝集成,确保高效的模型开发和部署。
我可以看一些来自 LVIS 数据集的示例注释吗?
是的,LVIS 数据集包含多种具有不同对象类别和复杂场景的图像。这里是一张示例图像及其注释:
这幅马赛克图像展示了一个训练批次,由多个数据集图像组合而成。马赛克增加了每个训练批次中对象和场景的多样性,增强了模型在不同环境下的泛化能力。有关 LVIS 数据集的更多详细信息,请查阅 LVIS 数据集文档。
COCO8 数据集
简介
Ultralytics COCO8 是一个小型但多用途的物体检测数据集,由 COCO train 2017 集的前 8 张图像组成,其中 4 张用于训练,4 张用于验证。此数据集非常适合测试和调试物体检测模型,或者尝试新的检测方法。由于只有 8 张图像,它非常易于管理,但又足够多样化,可以用于检查训练管道中的错误,并在训练更大数据集之前进行健全性检查。
www.youtube.com/embed/uDrn9QZJ2lk
观看: Ultralytics COCO 数据集概述
该数据集适用于 Ultralytics 的HUB和YOLOv8使用。
数据集 YAML
YAML(另一种标记语言)文件用于定义数据集配置。它包含有关数据集路径、类别和其他相关信息。对于 COCO8 数据集,coco8.yaml
文件位于github.com/ultralytics/ultralytics/blob/main/ultralytics/cfg/datasets/coco8.yaml
。
ultralytics/cfg/datasets/coco8.yaml
# Ultralytics YOLO 🚀, AGPL-3.0 license
# COCO8 dataset (first 8 images from COCO train2017) by Ultralytics
# Documentation: https://docs.ultralytics.com/datasets/detect/coco8/
# Example usage: yolo train data=coco8.yaml
# parent
# ├── ultralytics
# └── datasets
# └── coco8 ← downloads here (1 MB)
# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
path: ../datasets/coco8 # dataset root dir
train: images/train # train images (relative to 'path') 4 images
val: images/val # val images (relative to 'path') 4 images
test: # test images (optional)
# Classes
names:
0: person
1: bicycle
2: car
3: motorcycle
4: airplane
5: bus
6: train
7: truck
8: boat
9: traffic light
10: fire hydrant
11: stop sign
12: parking meter
13: bench
14: bird
15: cat
16: dog
17: horse
18: sheep
19: cow
20: elephant
21: bear
22: zebra
23: giraffe
24: backpack
25: umbrella
26: handbag
27: tie
28: suitcase
29: frisbee
30: skis
31: snowboard
32: sports ball
33: kite
34: baseball bat
35: baseball glove
36: skateboard
37: surfboard
38: tennis racket
39: bottle
40: wine glass
41: cup
42: fork
43: knife
44: spoon
45: bowl
46: banana
47: apple
48: sandwich
49: orange
50: broccoli
51: carrot
52: hot dog
53: pizza
54: donut
55: cake
56: chair
57: couch
58: potted plant
59: bed
60: dining table
61: toilet
62: tv
63: laptop
64: mouse
65: remote
66: keyboard
67: cell phone
68: microwave
69: oven
70: toaster
71: sink
72: refrigerator
73: book
74: clock
75: vase
76: scissors
77: teddy bear
78: hair drier
79: toothbrush
# Download script/URL (optional)
download: https://github.com/ultralytics/assets/releases/download/v0.0.0/coco8.zip
使用
要在 COCO8 数据集上训练一个 YOLOv8n 模型,使用 640 的图像大小进行 100 个 epoch,您可以使用以下代码片段。有关可用参数的详细列表,请参考模型训练页面。
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="coco8.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=coco8.yaml model=yolov8n.pt epochs=100 imgsz=640
样本图像和注释
下面是 COCO8 数据集中一些图像的示例,以及它们相应的注释:
- 马赛克图像: 此图展示了由马赛克数据集图像组成的训练批次。马赛克是一种在训练过程中使用的技术,将多个图像组合成单个图像,以增加每个训练批次中对象和场景的多样性。这有助于提高模型对不同对象大小、长宽比和背景情境的泛化能力。
该示例展示了 COCO8 数据集图像的多样性和复杂性,以及在训练过程中使用马赛克的好处。
引用和致谢
如果您在研究或开发工作中使用 COCO 数据集,请引用以下论文:
@misc{lin2015microsoft,
title={Microsoft COCO: Common Objects in Context},
author={Tsung-Yi Lin and Michael Maire and Serge Belongie and Lubomir Bourdev and Ross Girshick and James Hays and Pietro Perona and Deva Ramanan and C. Lawrence Zitnick and Piotr Dollár},
year={2015},
eprint={1405.0312},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
我们要感谢 COCO 联盟为计算机视觉社区创建和维护这一宝贵资源。有关 COCO 数据集及其创建者的更多信息,请访问COCO 数据集网站。
常见问题解答
Ultralytics COCO8 数据集用于什么?
Ultralytics COCO8 数据集是一个紧凑而多功能的目标检测数据集,包括来自 COCO 2017 训练集的前 8 张图像,其中有 4 张用于训练,4 张用于验证。它旨在用于测试和调试目标检测模型,以及尝试新的检测方法。尽管规模较小,COCO8 提供了足够的多样性,可用作在部署更大数据集之前对训练流水线进行验收测试。详细信息请查看COCO8 数据集。
如何使用 COCO8 数据集训练 YOLOv8 模型?
要在 COCO8 数据集上训练 YOLOv8 模型,您可以使用 Python 或 CLI 命令。以下是如何开始的方式:
训练示例
from ultralytics import YOLO
# Load a model
model = YOLO("yolov8n.pt") # load a pretrained model (recommended for training)
# Train the model
results = model.train(data="coco8.yaml", epochs=100, imgsz=640)
# Start training from a pretrained *.pt model
yolo detect train data=coco8.yaml model=yolov8n.pt epochs=100 imgsz=640
欲获取所有可用参数的详尽列表,请参阅模型训练页面。
为何应使用 Ultralytics HUB 管理我的 COCO8 训练?
Ultralytics HUB 是一个全方位的网络工具,旨在简化 YOLO 模型的训练和部署,包括 Ultralytics YOLOv8 模型在 COCO8 数据集上的应用。它提供云端训练、实时跟踪和无缝数据集管理。HUB 允许您一键启动训练,避免手动设置的复杂性。了解更多关于Ultralytics HUB及其优势。
在使用 COCO8 数据集进行训练时,采用马赛克增强有什么好处?
在 COCO8 数据集中演示的马赛克增强技术,在训练期间将多个图像合并成单个图像。此技术增加了每个训练批次中对象和场景的多样性,提高了模型在不同对象大小、长宽比和场景背景下的泛化能力。从而形成更强大的目标检测模型。详细信息请参阅训练指南。
如何验证在 COCO8 数据集上训练的 YOLOv8 模型?
使用模型的验证命令,可以验证在 COCO8 数据集上训练的 YOLOv8 模型。您可以通过 CLI 或 Python 脚本调用验证模式,评估模型在精确指标下的性能。详细指南请访问验证页面。