OpenCV(4.10.0) d:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'cv::normalizeAnchor' while using img2table

[GiriMadhav722001]

I am getting above error while using img2table, please help me to fix the issue
below is the code which I used,

`from img2table.ocr import PaddleOCR
from img2table.document import Image

ocr = PaddleOCR(lang="en")
doc=Image('/content/cropped_image.png')

extracted_tables = doc.extract_tables(ocr=ocr,
implicit_rows=False,
implicit_columns=False,
borderless_tables=True,
min_confidence=50)`

[GiriMadhav722001]

please help me

[hbh112233abc]

when min_line_length < 6 then kernel_dims will as (1,0) or (0,1) but cv2.getStructuringElement params kernel_dims values must > 0 , my solution is change round to ceil function, then kernel_dims values min value = 1,and not raise above exception again

[GiriMadhav722001]

still i am getting same error
/usr/local/lib/python3.10/dist-packages/img2table/tables/processing/bordered_tables/lines.py in identify_straight_lines(thresh, min_line_length, char_length, vertical)
21 # Apply masking on image
22 kernel_dims = (1, round(min_line_length / 3)) if vertical else (round(min_line_length / 3), 1)
---> 23 kernel = cv2.getStructuringElement(cv2.MORPH_RECT, kernel_dims)
24 mask = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=1)
25

error: OpenCV(4.10.0) /io/opencv/modules/imgproc/src/filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'normalizeAnchor'

[GiriMadhav722001]

please help me, exception raised only when I am passing YOLO detected image having table instead of whole pdf,

[GiriMadhav722001]

cannot recognize vertically reversed text, is not recognized
see the below pictures this is actually not recognized and also some symbols like '-' not recognized

[GiriMadhav722001]

ocr=PaddleOCR(lang='en',use_angle_cls=True)
this extracts all text data from table even text is reversed vertically, extracts symbols as well
when I tried to use this above code I got below error as, please allow use to use that 'use_angle_cls=True'

TypeError Traceback (most recent call last)
in <cell line: 5>()
3
4 # Instantiation of OCR
----> 5 ocr = PaddleOCR(lang="en",use_angle_cls=True)
6 # ocr = EasyOCR(lang=["en"], kw={"gpu": False}
7 # ocr=paddle_ocr = PaddleOCR(lang="en", kw={"use_dilation": True})

TypeError: PaddleOCR.init() got an unexpected keyword argument 'use_angle_cls'

[GiriMadhav722001]

please help me please

[hbh112233abc]

ocr=PaddleOCR(lang='en',use_angle_cls=True) this extracts all text data from table even text is reversed vertically, extracts symbols as well when I tried to use this above code I got below error as, please allow use to use that 'use_angle_cls=True'

TypeError Traceback (most recent call last) in <cell line: 5>() 3 4 # Instantiation of OCR ----> 5 ocr = PaddleOCR(lang="en",use_angle_cls=True) 6 # ocr = EasyOCR(lang=["en"], kw={"gpu": False} 7 # ocr=paddle_ocr = PaddleOCR(lang="en", kw={"use_dilation": True})

TypeError: PaddleOCR.init() got an unexpected keyword argument 'use_angle_cls'

look source code img2table.ocr.paddle.py PaddleOCR only parse tow params lang and kw(dict), so init PaddleOCR you should write like below:

ocr = PaddleOCR(lang='en',kw={'use_angle_cls':True})

the kw dict should like baidu's PaddleOCR class params

[GiriMadhav722001]

still i am getting same error
/usr/local/lib/python3.10/dist-packages/img2table/tables/processing/bordered_tables/lines.py in identify_straight_lines(thresh, min_line_length, char_length, vertical)
21 # Apply masking on image
22 kernel_dims = (1, round(min_line_length / 3)) if vertical else (round(min_line_length / 3), 1)
---> 23 kernel = cv2.getStructuringElement(cv2.MORPH_RECT, kernel_dims)
24 mask = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=1)
25

error: OpenCV(4.10.0) /io/opencv/modules/imgproc/src/filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'normalizeAnchor'

please provide me a solution for this as well

[GiriMadhavB]

Some time randomly OpenCV(4.10.0) d:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'cv::normalizeAnchor' while using img2table this exception has raised how to address this please help me

[GiriMadhav722001]

Some time randomly OpenCV(4.10.0) d:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'cv::normalizeAnchor'
This error occurs randomly; how can I address it

[GiriMadhav722001]

please help me

[kirillkoncha]

downgrade to 1.2.11 solved this problem for me

[GiriMadhav722001]

how to install this still showing 1.3.0
even i was used this
!pip install img2table==1.2.
!pip install img2table[paddle]==1.2.11

[GiriMadhav722001]

friends please help me

for the above image i am getting the error
error Traceback (most recent call last)
Cell In[251], line 3
1 doc = Image(r"C:\Users\617468422\Desktop\tabledetction\tableDetectionDataExtractiontoExcel\output\detectedTables\4737739_1060182_ARN004_55841_PO0196_M004_EE_AS_AB2\page_13_table0_expanded_top.png")
2 # # Extract tables
----> 3 extracted_tables = doc.extract_tables(ocr=pocr,
4 implicit_rows=False,
5 implicit_columns=False,
6 borderless_tables=False,
7 min_confidence=50)

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\document\image.py:48, in Image.extract_tables(self, ocr, implicit_rows, implicit_columns, borderless_tables, min_confidence)
37 def extract_tables(self, ocr: "OCRInstance" = None, implicit_rows: bool = False, implicit_columns: bool = False,
38 borderless_tables: bool = False, min_confidence: int = 50) -> List[ExtractedTable]:
39 """
40 Extract tables from document
41 :param ocr: OCRInstance object used to extract table content
(...)
46 :return: list of extracted tables
47 """
---> 48 extracted_tables = super(Image, self).extract_tables(ocr=ocr,
49 implicit_rows=implicit_rows,
50 implicit_columns=implicit_columns,
51 borderless_tables=borderless_tables,
52 min_confidence=min_confidence)
53 return extracted_tables.get(0)

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\document\base_init_.py:128, in Document.extract_tables(self, ocr, implicit_rows, implicit_columns, borderless_tables, min_confidence)
125 # Extract tables from document
126 from img2table.tables.image import TableImage
127 tables = {idx: TableImage(img=img,
--> 128 min_confidence=min_confidence).extract_tables(implicit_rows=implicit_rows,
129 implicit_columns=implicit_columns,
130 borderless_tables=borderless_tables)
131 for idx, img in enumerate(self.images)}
133 # Update table content with OCR if possible
134 tables = self.get_table_content(tables=tables,
135 ocr=ocr,
136 min_confidence=min_confidence)

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\tables\image.py:129, in TableImage.extract_tables(self, implicit_rows, implicit_columns, borderless_tables)
126 return []
128 # Extract bordered tables
--> 129 self.extract_bordered_tables(implicit_rows=implicit_rows,
130 implicit_columns=implicit_columns)
132 if borderless_tables:
133 # Extract borderless tables
134 self.extract_borderless_tables()

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\tables\image.py:66, in TableImage.extract_bordered_tables(self, implicit_rows, implicit_columns)
63 min_line_length = int(min(1.5 * self.median_line_sep, 4 * self.char_length)) if self.median_line_sep else 20
65 # Detect rows in image
---> 66 h_lines, v_lines = detect_lines(img=self.img,
67 contours=self.contours,
68 char_length=self.char_length,
69 min_line_length=min_line_length)
70 self.lines = h_lines + v_lines
72 # Create cells from rows

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\tables\processing\bordered_tables\lines.py:112, in detect_lines(img, contours, char_length, min_line_length)
109 binary_img = 255 * (edge_img >= min(2.5 * np.mean(edge_img), np.max(edge_img))).astype(np.uint8)
111 # Detect lines
--> 112 h_lines = identify_straight_lines(thresh=binary_img,
113 min_line_length=min_line_length,
114 char_length=char_length,
115 vertical=False)
116 v_lines = identify_straight_lines(thresh=binary_img,
117 min_line_length=min_line_length,
118 char_length=char_length,
119 vertical=True)
121 return h_lines, v_lines

File ~\AppData\Roaming\Python\Python312\site-packages\img2table\tables\processing\bordered_tables\lines.py:23, in identify_straight_lines(thresh, min_line_length, char_length, vertical)
21 # Apply masking on image
22 kernel_dims = (1, round(min_line_length / 3)) if vertical else (round(min_line_length / 3), 1)
---> 23 kernel = cv2.getStructuringElement(cv2.MORPH_RECT, kernel_dims)
24 mask = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=1)
26 # Apply closing for hollow lines

error: OpenCV(4.10.0) d:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\filterengine.hpp:367: error: (-215:Assertion failed) anchor.inside(Rect(0, 0, ksize.width, ksize.height)) in function 'cv::normalizeAnchor'