Giter Site home page Giter Site logo

paddlecv-sig / eiseg Goto Github PK

View Code? Open in Web Editor NEW
333.0 7.0 51.0 9.81 MB

高效交互式语义分割标注软件EISeg 『Efficient and intelligent interactive segmentation annotation software』

Home Page: https://github.com/PaddleCV-SIG/EISeg/wiki

License: Apache License 2.0

Python 90.80% Shell 0.10% C++ 3.39% C 2.73% Cython 2.98%
interactive-segmentation segmentation

eiseg's People

Contributors

fjjrichard avatar geoyee avatar gt-zhangacer avatar haoyuying avatar linhandev avatar michaelowenliu avatar yiakwy avatar youssef-harby avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar

eiseg's Issues

标签覆盖问题

目前实现是按照标注顺序,后面的盖前面的,考虑要不要实现按照标签序号顺序,大号的盖小号,或者反过来

win下安装报错

(p2) PS H:\EISeg> python -m eiseg
AttributeError: __getattribute__

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "F:\Users\livingbody\miniconda3\envs\p2\lib\runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "F:\Users\livingbody\miniconda3\envs\p2\lib\runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "H:\EISeg\eiseg\__main__.py", line 4, in <module>
    main()
  File "H:\EISeg\eiseg\run.py", line 21, in main
    window = APP_EISeg()  # 创建对象
  File "H:\EISeg\eiseg\app.py", line 50, in __init__
    self.setupUi(self)
  File "H:\EISeg\eiseg\ui.py", line 256, in setupUi
    QtCore.QMetaObject.connectSlotsByName(MainWindow)
SystemError: <built-in function connectSlotsByName> returned a result with an error set

模型输入point的组织形式

请问作者,对于模型前向传播时,输入的point参数的各个维度代表的含义可以稍微解释下吗?

比如,我点了一个点,输入的point的形状是:(2, 2, 3) , 我现在知道了 最后一个维度中包含的是 (y, x, index), 也知道 第二个维度的值是 正样本点的数量的2倍,但是我没看懂为啥第二个维度的值要是2,另外一部分存的是啥啊?(看结果,y值没变,x值变化)
第一维,嘿嘿也没看懂? 为啥是2,这里的含义是什么呢?

感谢您的精彩工作!

如何标注相邻的两个物体

比如上衣和裤子,标注出上衣后标裤子的时候上衣区域不能再标为负样本点,裤子的整体轮廓会包含上衣点集

模型下载

请问可以下载HRNet18_OCR64模型吗(后缀名为,pdmodel格式)

有关有遮挡物体被划分为多个bbox的问题。

作者您好,感谢您精彩的工作!

我在使用过程中发现,如果“待标注的物体”被其他“前景物体”从中间遮挡,而被分为两部分时,此时对“待标注的物体”进行标注时,会生成多个bbox,在json中也是被认为成“两个物体”而不是同一个物体,请问这个问题有办法解决吗?

期待您的回复!

重名覆盖

两张文件名相同,只有格式不同的图片在保存的时候标签都会存成 文件名.png,会互相覆盖

删除多边形偶见闪退

如题,闪退没有任何报错,这个问题偶见,不知道如何触发的。需要进一步收集关于这个问题的信息。

保留最大联通区问题

  1. 打开保留最大联通区后,不管选择哪个颜色的标签,交互点击对应的可视化mask都是第一个标签的蓝色。
  2. 当存在孔洞无法闭合时,保留最大联通区报错:
  File "e:\PdCVSIG\github\EISeg\eiseg\util\polygon.py", line 38, in get_polygon
    polygons[j].append(polygons[j][0])  # 闭合
AttributeError: 'NoneType' object has no attribute 'append'

eiseg COCO 格式用于 paddlex、PaddleDetection 训练报错

bug描述
使用 EISEG 标注后保存为 COCO 格式,用于paddlex、PaddleDetection 训练时会报错,请问是需要特别转换么?

[10/23 17:12:10] ppdet.data.source.coco WARNING: Found an invalid bbox in annotations: im_id: 346, area: 0.0 x1: 664.0, y1: 690.0, x2: 853.0, y2: 908.0.
Traceback (most recent call last):
  File "tools/train.py", line 138, in <module>
    main()
  File "tools/train.py", line 134, in main
    run(FLAGS, cfg)
  File "tools/train.py", line 100, in run
    trainer = Trainer(cfg, mode='train')
  File "/opt/deeplearning/PaddleDetection/ppdet/engine/trainer.py", line 74, in __init__
    self.loader = create('{}Reader'.format(self.mode.capitalize()))(
  File "/opt/deeplearning/PaddleDetection/ppdet/data/reader.py", line 163, in __call__
    self.dataset.parse_dataset()
  File "/opt/deeplearning/PaddleDetection/ppdet/data/source/coco.py", line 243, in parse_dataset
    assert ct > 0, 'not found any coco record in %s' % (anno_path)
AssertionError: not found any coco record in dataset/splotch_coco/train.json

运行环境(请尽量填写,这可以帮助我们定位问题):

  • 系统: Linux
  • 安装方式:pip
  • 软件版本:0.3.0.2

边缘重合问题

标注时有时希望是无缝的,是否需要链接边界重合,如何做较好?
image

为什么模型训练用两个优化器

optimizer1 = opt(learning_rate=lr(float(cfg.get('learning_rate').get('value_1')),
decay_steps=cfg.get('learning_rate').get('decay').get('steps'),
end_lr=cfg.get('learning_rate').get('decay').get('end_lr'),
power=cfg.get('learning_rate').get('decay').get('power')),
parameters=other_params)
optimizer2 = opt(learning_rate=lr(float(cfg.get('learning_rate').get('value_1')),
decay_steps=cfg.get('learning_rate').get('decay').get('steps'),
end_lr=cfg.get('learning_rate').get('decay').get('end_lr'),
power=cfg.get('learning_rate').get('decay').get('power')),
parameters=backbone_params)

undo/redo 图标

感觉undo/redo图标如果换成那种word里常见的,一个曲线,一个箭头,可能更直观一些

比如这种
image

是否提供高精度模型的训练代码?

看了eiseg的readme,以及所有issues,好像在25天前作者提到会近期开源相关训练代码,
但readme上还是没有见到相关代码。
目前想基于工业界的某一细分领域的数据集训练一个模型去做 后续数据集的标注,以提高标注质量和速度,
是否提供 hrnet18_ocr64_cocolvis 这个高精度模型的训练代码?谢谢!

Train module problem: ritm_train.py import error.

THX for sharing!
Import error was found in ritm_train.py
from model.model import (
get_hrnet_model,
DistMapsHRNetModel,
get_deeplab_model,
get_shufflenet_model,
)
these modules were not exist in model.py, even in this project. so where can i find these modules, please?

Could give *.ui file to update GUI?

For example, There is no .ui file to update GUI design in "/contrib/EIseg/", So I cannot add a button in this program.Thank you very much!

各位大牛们,能不能提供一下EIseg的UI文件啊,想往上加个按钮,求求了,谢谢大家!

Segmentation fault when running

Hi,
when I try this great label tool, I encounted ​such error and the Error Message Summary is as below:

Error Message Summary:

FatalError: Segmentation fault is detected by the operating system.
​[TimeInfo: *** Aborted at 1632232668 (unix time) try "date -d @1632232668" if you are using GNU date ***]
​[SignalInfo: *** SIGSEGV (@0x12) received by PID 7565 (TID 0x7f81cf5fb740) from PID 18 ***]

I use the cpu-version paddle paddle. What dose this error mean and how to avoid it?
Best.

多边形可选择问题

能否设置一键开启或关闭多边形标注,或者能够对单独的mask进行多边形化

正点过后的第一个负点无效

点击正点可以正常的扩充mask,但目前版本点击的第一个负点对mask的起到的作用微乎其微,而点击第二个负点后第一个负点的作用才开始体现。可能是显示的问题。需要查明和修改

直接打开图像报错

直接打开一张没有标签的图像报错:

  File "e:\PdCVSIG\github\EISeg\eiseg\app.py", line 1092, in loadLabel
    imgId = self.coco.imgNameToId.get(osp.basename(imgPath), None)
AttributeError: 'NoneType' object has no attribute 'imgNameToId'

大图加载问题

超大的图打开直接切分宫格,不预先加载,造成未响应时间过长

coco格式

image
您好,我想请问,保存出来的coco文件中的bbox是什么意思啊?

json

标注保存的json文件和coco数据集json文件格式有什么区别,

EISEG的菜单栏功能

图片
之前通过pycharm运行eiseg时菜单栏有许多功能,比如选择是否适用于遥感图像,但是现在打开eiseg时菜单栏只剩下文件和标注这两个功能,请问如何运行回到原来的状态?

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.