SLANeXt_wireless

Introduction

Table structure recognition is an important component of table recognition systems, capable of converting non-editable table images into editable table formats (such as HTML). The goal of table structure recognition is to identify the positions of rows, columns, and cells in tables. The performance of this module directly affects the accuracy and efficiency of the entire table recognition system. The table structure recognition module usually outputs HTML code for the table area, which is then passed as input to the tabl recognition pipeline for further processing.

Note: The accuracy of SLANeXt_wireless comes from the results of joint testing with SLANeXt_wired.

Installation

1. PaddlePaddle

Please refer to the following commands to install PaddlePaddle using pip:

# for CUDA11.8
python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/

# for CUDA12.6
python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/

# for CPU
python -m pip install paddlepaddle==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/

For details about PaddlePaddle installation, please refer to the PaddlePaddle official website.

2. PaddleOCR

Install the latest version of the PaddleOCR inference package from PyPI:

python -m pip install paddleocr

Model Usage

You can quickly experience the functionality with a single command:

paddleocr table_structure_recognition \
    --model_name SLANeXt_wireless \
    -i https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/6rfhb-CXOHowonjpBsaUJ.png

You can also integrate the model inference of the table classification module into your project. Before running the following code, please download the sample image to your local machine.

from paddleocr import TableStructureRecognition
model = TableStructureRecognition(model_name="SLANeXt_wireless")
output = model.predict(input="6rfhb-CXOHowonjpBsaUJ.png", batch_size=1)
for res in output:
    res.print(json_format=False)
    res.save_to_json("./output/res.json")

After running, the obtained result is as follows:

{'res': {'input_path': '6rfhb-CXOHowonjpBsaUJ.png', 'page_index': None, 'bbox': [[5, 4, 48, 5, 46, 85, 5, 81], [84, 6, 146, 6, 143, 101, 83, 98], [186, 6, 217, 6, 212, 104, 184, 98], [239, 7, 281, 8, 276, 107, 235, 108], [324, 6, 405, 6, 404, 105, 323, 106], [405, 4, 488, 5, 488, 100, 403, 94], [3, 56, 96, 60, 95, 187, 3, 180], [108, 68, 157, 71, 159, 193, 110, 187], [179, 75, 207, 79, 211, 199, 184, 192], [238, 72, 277, 76, 281, 203, 243, 199], [318, 68, 400, 70, 404, 207, 325, 205], [395, 66, 494, 68, 494, 214, 397, 212], [11, 138, 62, 145, 68, 329, 12, 321], [105, 151, 156, 158, 171, 332, 117, 323], [177, 157, 210, 166, 229, 322, 197, 312], [232, 152, 276, 159, 295, 322, 253, 316], [313, 142, 396, 147, 409, 330, 332, 326], [392, 139, 491, 144, 492, 332, 404, 330], [3, 239, 86, 254, 103, 450, 3, 445], [97, 251, 152, 261, 176, 458, 116, 454], [172, 254, 211, 265, 239, 461, 200, 458], [235, 248, 289, 257, 316, 466, 264, 464], [310, 235, 402, 242, 419, 469, 337, 468], [381, 229, 491, 236, 492, 469, 400, 468], [9, 340, 74, 361, 88, 490, 11, 489], [95, 338, 129, 353, 150, 493, 113, 492], [176, 342, 192, 358, 221, 493, 206, 492], [235, 335, 261, 351, 289, 493, 265, 492], [310, 325, 372, 339, 393, 493, 338, 493], [382, 321, 482, 334, 485, 493, 402, 493]], 'structure': ['', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', '', ''], 'structure_score': 0.9999998}}

For details about usage command and descriptions of parameters, please refer to the Document.

Pipeline Usage

The ability of a single model is limited. But the pipeline consists of several models can provide more capacity to resolve difficult problems in real-world scenarios.

General Table Recognition V2 Pipeline

The general table recognition V2 pipeline is used to solve table recognition tasks by extracting information from images and outputting it in HTML or Excel format. And there are 8 modules in the pipeline:

• Table Classification Module
• Table Structure Recognition Module
• Table Cell Detection Module
• Text Detection Module
• Text Recognition Module
• Layout Region Detection Module (Optional)
• Document Image Orientation Classification Module (Optional)
• Text Image Unwarping Module (Optional)

Run a single command to quickly experience the general table recognition V2 pipeline:

paddleocr table_recognition_v2 -i https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/mabagznApI1k9R8qFoTLc.png  \
    --use_doc_orientation_classify False  \
    --use_doc_unwarping False \
    --save_path ./output \
    --device gpu:0

Results are printed to the terminal:

{'res': {'input_path': 'mabagznApI1k9R8qFoTLc.png', 'page_index': None, 'model_settings': {'use_doc_preprocessor': False, 'use_layout_detection': True, 'use_ocr_model': True}, 'layout_det_res': {'input_path': None, 'page_index': None, 'boxes': [{'cls_id': 8, 'label': 'table', 'score': 0.86655592918396, 'coordinate': [0.0125130415, 0.41920784, 1281.3737, 585.3884]}]}, 'overall_ocr_res': {'input_path': None, 'page_index': None, 'model_settings': {'use_doc_preprocessor': False, 'use_textline_orientation': False}, 'dt_polys': array([[[   9,   21],
        ...,
        [   9,   59]],

       ...,

       [[1046,  536],
        ...,
        [1046,  573]]], dtype=int16), 'text_det_params': {'limit_side_len': 960, 'limit_type': 'max', 'thresh': 0.3, 'box_thresh': 0.6, 'unclip_ratio': 2.0}, 'text_type': 'general', 'textline_orientation_angles': array([