evezerest / ppocrlabel Goto Github PK

当前PPOCRLabel 点“保存识别结果” 后会将所有图片识别出来的文本保存到一个 rec_gt.txt 文件中，请问能不能设置为各个图片识别出来的文本进行单独保存呢？谢谢
例如：001.jpg、002.jpg、003.jpg、004.jpg、005.jpg
识别结果保存为：001.txt、002.txt、003.txt、004.txt、005.txt

CANNOT OPEN .xlsx. It could be one of the following reasons: Only support Windows | No python win32com
CANNOT OPEN .xlsx. It could be the following reasons: .xlsx is not existed

环境：

Windows10
Python 3.8
CUDA Toolkit 11.2.2
CUDA 12.1
pyqt5 5.15.9

错误信息：

initPos PyQt5.QtCore.QPointF(1018.2962962962963, 119.0)
result in reRec is [[[1018, 119], [1214, 119], [1214, 163], [1018, 163]], [('S1.00', 0.9850572347640991)]]
Traceback (most recent call last):
File "PPOCRLabel.py", line 2139, in singleRerecognition
self.singleLabel(shape)
File "PPOCRLabel.py", line 1112, in singleLabel
item.setText(shape.label)
TypeError: setText(self, atext: str): argument 1 has unexpected type 'tuple'

How to remove PPOCRLabel

如题

ubuntu下不能表格标注吗

iImportError: cannot import name 'check_and_read' from 'ppocr.utils.utility'

目前我使用100*30的提示说太小无法识别

经常闪退怎么处理

Please update requirstments.txt PPOCRLabel is not work with newest some packages anymore.
for instance:
PyQt needs to be 5.14 newest is 5.15
and onnx should be 1.12

enviroment: Ubuntu 20.04, Python 3.7

1. pip install paddlepaddle
2. pip install PPOCRLabel
3. sudo apt-get install libxcb-xinerama0
4. pip uninstall opencv-python
5. pip install opencv-python-headless

折腾了半天，就是需要这样安装。至于opencv-python-headless 的版本，最新的我这里可以，但也可以用4.2.0.32
至于 opencv-contrib-headless 需不需要安装，我这里是不需要安装。但是绝对不能安装 opencv-python。

I had been using PPOCRLabel to create labels for layout detection. The output of PPOCRLabel is a label.txt file that has its own format. I need to know if can I use this label.txt file directly to train a layout model (in PPStructure) or do I need to convert it to COCO annotations format. Is there an available option to convert it to COCO annotations (.json) format?

Thank you

`
(paddle) root@ubuntu-3090ti:/home/ubuntu/github/PaddleOCR/PPOCRLabel# python PPOCRLabel.py --lang ch
Traceback (most recent call last):
File "/home/ubuntu/github/PaddleOCR/PPOCRLabel/PPOCRLabel.py", line 27, in
from PyQt5.QtCore import QSize, Qt, QPoint, QByteArray, QTimer, QFileInfo, QPointF, QProcess
ImportError: /root/anaconda3/envs/paddle/lib/python3.9/site-packages/PyQt5/QtCore.abi3.so: undefined symbol: _ZdaPvm, version Qt_5

Linux ubuntu-3090ti 5.15.0-48-generic #54-Ubuntu SMP Fri Aug 26 13:26:29 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
`

When scrolling:

Traceback (most recent call last):
File "C:\Users\thomas\AppData\Local\Programs\Python\Python310\lib\site-packages\PPOCRLabel\PPOCRLabel.py", line 1425, in scrollRequest
bar.setValue(bar.value() + bar.singleStep() * units)
TypeError: setValue(self, a0: int): argument 1 has unexpected type 'float'

When trying to draw RectBox:

Traceback (most recent call last):
File "C:\Users\thomas\AppData\Local\Programs\Python\Python310\lib\site-packages\PPOCRLabel\libs\canvas.py", line 596, in paintEvent
p.drawRect(leftTop.x(), leftTop.y(), rectWidth, rectHeight)
TypeError: arguments did not match any overloaded call:
drawRect(self, rect: QRectF): argument 1 has unexpected type 'float'
drawRect(self, x: int, y: int, w: int, h: int): argument 1 has unexpected type 'float'
drawRect(self, r: QRect): argument 1 has unexpected type 'float'

Hi,

I got this error when try to launch PPOCRLabel:

$ PPOCRLabel
QObject::moveToThread: Current thread (0x213f630) is not the object's thread (0x51f3a50).
Cannot move to target thread (0x213f630)

qt.qpa.plugin: Could not load the Qt platform plugin "xcb" in "/home/nicola/miniconda3/envs/head-less/lib/python3.11/site-packages/cv2/qt/plugins" even though it was found.
This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.

Available platform plugins are: xcb, eglfs, linuxfb, minimal, minimalegl, offscreen, vnc, wayland-egl, wayland, wayland-xcomposite-egl, wayland-xcomposite-glx, webgl.

Any idea?

Thank you.

代码地址：https://blog.csdn.net/qq_37668436/article/details/109850815

os: Macos BigSur 11.7
Paddle： 2.32, PaddleOCR： 2.6.1.0, PPOCRLabel: 2.1.1, PyQt5 : 5.15.7, PyQt5-Qt5: 5.15.2, PyQt5-sip: 12.11.0

File "/opt/anaconda3/envs/pythonProject/lib/python3.8/site-packages/PPOCRLabel/libs/hashableQListWidgetItem.py", line 37, in __init__
    super(HashableQListWidgetItem, self).__init__(*args)
TypeError: arguments did not match any overloaded call:
  QListWidgetItem(parent: QListWidget = None, type: int = QListWidgetItem.Type): argument 1 has unexpected type 'tuple'
  QListWidgetItem(str, parent: QListWidget = None, type: int = QListWidgetItem.Type): argument 1 has unexpected type 'tuple'
  QListWidgetItem(QIcon, str, parent: QListWidget = None, type: int = QListWidgetItem.Type): argument 1 has unexpected type 'tuple'
  QListWidgetItem(QListWidgetItem): argument 1 has unexpected type 'tuple'
zsh: abort      PPOCRLabel

I am trying to annotate a table using PPOcrLabel. Following the instructions here (https://github.com/Evezerest/PPOCRLabel/tree/release2.0?tab=readme-ov-file#22-table-annotation ). But when I change the value of any cell and run Cell Re-Recognition. I am getting the following error and application crashes.

Traceback (most recent call last):
File "f:\pycharmprojects\ppstructurelearning\venv\lib\site-packages\PPOCRLabel\PPOCRLabel.py", line 2424, in cellreRecognition
bboxes = self.ocr.ocr(img_crop, det=True, rec=False, cls=False)[0]
File "f:\pycharmprojects\ppstructurelearning\venv\lib\site-packages\paddleocr\paddleocr.py", line 674, in ocr
if not dt_boxes:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Please help !!

Hey team,

We found an issue here:

Say if we have an image and three bounding boxes A, B and C, and save them by Export Recognition Result. Then after a while, we notice A is being mislabeled. So we deleted A inside the app. However, if we hit Export Recognition Result again, we'll notice the cropped image of A is still there in the file system.

Any thoughts in removing deleted bounding boxes in the file system?

Can i use another language for text annotation (for example vietnamese)?
Thank you for this useful tool.

Please support rotated rect like powerpoint

I notice each detection box is associated with a "difficult" field in the Label.txt, but I didn't find any way to modify this field in the UI, how can we achieve this?

Furthermore, is it possible for us to customize this field? For example, I wanted to add a "logo" field there to denote whether the content inside the bounding box is a logo, any easy way to achieve this?

how to edit label.txt in ppocrlabel tool..

I dont find any option to load the label.txt

Hello , when i try to use rect box i have bug like
AppData\Local\Programs\Python\Python310\lib\site-packages\PPOCRLabel\libs\canvas.py", line 596, in paintEvent
p.drawRect(leftTop.x(), leftTop.y(), rectWidth, rectHeight)
TypeError: arguments did not match any overloaded call:
drawRect(self, QRectF): argument 1 has unexpected type 'float'
drawRect(self, int, int, int, int): argument 1 has unexpected type 'float'
drawRect(self, QRect): argument 1 has unexpected type 'float'
What should i do to fix this bug ?