All Projects → hzylmf → Od Annotation

hzylmf / Od Annotation

目标检测数据集标注工具

Programming Languages

python
139335 projects - #7 most used programming language
python3
1442 projects

Projects that are alternatives of or similar to Od Annotation

Myvision
Computer vision based ML training data generation tool 🚀
Stars: ✭ 453 (+79.05%)
Mutual labels:  object-detection, annotation
Diffgram
Data Annotation, Data Labeling, Annotation Tooling, Training Data for Machine Learning
Stars: ✭ 43 (-83%)
Mutual labels:  object-detection, annotation
Yolo label
GUI for marking bounded boxes of objects in images for training neural network Yolo v3 and v2 https://github.com/AlexeyAB/darknet, https://github.com/pjreddie/darknet
Stars: ✭ 128 (-49.41%)
Mutual labels:  object-detection, annotation
Nanodet
⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥
Stars: ✭ 3,640 (+1338.74%)
Mutual labels:  object-detection
Ssd Variants
PyTorch implementation of several SSD based object detection algorithms.
Stars: ✭ 233 (-7.91%)
Mutual labels:  object-detection
Taco
🌮 Trash Annotations in Context Dataset Toolkit
Stars: ✭ 243 (-3.95%)
Mutual labels:  object-detection
Keras Retinanet For Open Images Challenge 2018
Code for 15th place in Kaggle Google AI Open Images - Object Detection Track
Stars: ✭ 251 (-0.79%)
Mutual labels:  object-detection
Flip
Synthetic Image generation with Flip. Generate thousands of new 2D images from a small batch of objects and backgrounds.
Stars: ✭ 231 (-8.7%)
Mutual labels:  object-detection
Computer Vision Guide
📖 This guide is to help you understand the basics of the computerized image and develop computer vision projects with OpenCV. Includes Python, Java, JavaScript, C# and C++ examples.
Stars: ✭ 244 (-3.56%)
Mutual labels:  object-detection
Mxnet Yolo
YOLO: You only look once real-time object detector
Stars: ✭ 240 (-5.14%)
Mutual labels:  object-detection
Silencer
Scala compiler plugin for warning suppression
Stars: ✭ 239 (-5.53%)
Mutual labels:  annotation
D2det
D2Det: Towards High Quality Object Detection and Instance Segmentation (CVPR2020)
Stars: ✭ 234 (-7.51%)
Mutual labels:  object-detection
Real time object detection and tracking
YOLOv2 and MobileNet_SSD detection algorithms used along with KCF object tracker
Stars: ✭ 241 (-4.74%)
Mutual labels:  object-detection
Retinaface
The remake of the https://github.com/biubug6/Pytorch_Retinaface
Stars: ✭ 226 (-10.67%)
Mutual labels:  object-detection
Vovnet Detectron2
VoVNet backbone networks for detectron2, in CVPR 2020
Stars: ✭ 249 (-1.58%)
Mutual labels:  object-detection
Prodigy Recipes
🍳 Recipes for the Prodigy, our fully scriptable annotation tool
Stars: ✭ 229 (-9.49%)
Mutual labels:  annotation
Browser Extension
The Hypothesis browser extensions.
Stars: ✭ 246 (-2.77%)
Mutual labels:  annotation
Arkit2.0 Prototype
After Apple’s introduction of ARKit 2, we have been consistently working behind to create shared-AR experiences. Our goal is to improve the utility of mobile using AR experiences.
Stars: ✭ 236 (-6.72%)
Mutual labels:  object-detection
Fcos
FCOS: Fully Convolutional One-Stage Object Detection (ICCV'19)
Stars: ✭ 2,839 (+1022.13%)
Mutual labels:  object-detection
Yolobile
This is the implementation of YOLObile: Real-Time Object Detection on Mobile Devices via Compression-Compilation Co-Design
Stars: ✭ 201 (-20.55%)
Mutual labels:  object-detection

目标检测数据集标注工具

采用python-flask框架开发,基于B/S方式交互,支持多人同时标注。

(丑陋的)界面效果图

特点

  • B/S方式交互
  • 支持多人同时标注(可分配不同标注人员的标注范围,或不同人员标注不同类别)
  • 类别采用选择方式,免去手工输入类别工作
  • 支持拖拽方式修正标注区域
  • 支持键盘方向键切换标注样本
  • 支持多类别多目标标注

使用方法

  1. 根据requirements.txt安装环境依赖
$ cd od-annotation
$ pip3 install -r requirements.txt
  1. 重命名标注样本,采用前导0方式编号,共6位(000001-0000xx),注意保持样本编号连续。
  2. 编辑annotation/label_config.txt文件,根据格式配置标签
# 标签名称:标签描述
dog:狗
  1. 编辑config.py,根据样本实际情况修改:
SAMPLE_FILE_TYPE = 'jpg'  # 样本图片格式
SAMPLE_FILE_PATH = 'your samples directory path'  # 样本图片存放目录
  1. 启动/停止/重启标注工具:
$ cd od-annotation
$ python3 app.py --start|stop|restart  # 前台进程方式运行
$ python3 app.py --start|restart --daemon  # 以后台进程方式(重新)启动
  1. 访问http://localhost:5000开始标注。先选定待标注类别,然后按住鼠标左键并拖拽鼠标在目标区域绘制矩形框进行标注,松开鼠标完成标注。可拖动矩形框以修正坐标,右击某矩形框可取消该标注。每次新绘制矩形框、拖动已有矩形框或右击取消矩形框时,会在下方的当前样本标注状态文本框中同步更新该样本的标注结果。
  2. 点击左右方向按钮或通过键盘方向键切换标注样本。切换时自动提交标注结果,同时在所有样本标注状态文本框中更新当前样本的标注结果。或手动点击保存按钮提交标注结果。
  3. 所有样本标注完成后,若需要转换成VOC2007格式,执行:
$ cd od-annotation
$ python3 app.py --convert2voc

查看annotation/VOC2007目录下相关文件夹是否生成对应文件

说明

  • 依赖python3
  • Windows下python无fork功能,不支持daemon(后台进程)方式启动
  • 标注数据在annotation/annotation.txt文件中,每行一条标注数据,格式为filename,x1,y1,x2,y2,classname,x1,y1,x2,y2分别表示左上角和右下角坐标

已知Bug

  • 绘制区域再选择对应类别,然后切换样本时会导致类别单选框状态跟着切换(临时解决方法:通过点击页面空白区域来取消单选框焦点以避免bug)
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].