## 1. Introduction to Seal Text Recognition Pipeline
Seal text recognition is a technology that automatically extracts and recognizes the content of seals from documents or images. The recognition of seal text is part of document processing and has many applications in various scenarios, such as contract comparison, warehouse entry and exit review, and invoice reimbursement review.
The seal text recognition pipeline is used to recognize the text content of seals, extracting the text information from seal images and outputting it in text form. This pipeline integrates the industry-renowned end-to-end OCR system PP-OCRv4, supporting the detection and recognition of curved seal text. Additionally, this pipeline integrates an optional layout region localization module, which can accurately locate the layout position of the seal within the entire document. It also includes optional document image orientation correction and distortion correction functions. Based on this pipeline, millisecond-level accurate text content prediction can be achieved on a CPU. This pipeline also provides flexible service deployment methods, supporting the use of multiple programming languages on various hardware. Moreover, it offers custom development capabilities, allowing you to train and fine-tune on your own dataset based on this pipeline, and the trained model can be seamlessly integrated.
<b>The seal text recognition</b> pipeline includes a seal text detection module and a text recognition module, as well as optional layout detection module, document image orientation classification module, and text image correction module.
- [Seal Text Detection Module](../module_usage/seal_text_detection.en.md)
<td>A higher precision layout region localization model based on RT-DETR-L trained on a self-built dataset including Chinese and English papers, multi-column magazines, newspapers, PPTs, contracts, books, exam papers, research reports, ancient books, Japanese documents, and vertical text documents</td>
<td>A high precision layout region localization model based on RT-DETR-L trained on a self-built dataset including Chinese and English papers, magazines, contracts, books, exam papers, and research reports</td>
<td>A balanced model of accuracy and efficiency based on PicoDet-L trained on a self-built dataset including Chinese and English papers, magazines, contracts, books, exam papers, and research reports</td>
<td>A highly efficient layout region localization model based on PicoDet-S trained on a self-built dataset including Chinese and English papers, magazines, contracts, books, exam papers, and research reports</td>
>❗ Listed above are the <b>4 core models</b> that are the focus of the layout detection module, which supports a total of <b>13 full models</b>, including multiple models with pre-defined different categories, among which 9 models include the seal category. Apart from the 3 core models mentioned above, the remaining models are as follows:
<td>A highly efficient layout region localization model based on the lightweight PicoDet-S model trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<td>An efficiency-accuracy balanced layout region localization model based on PicoDet-L trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<td>A high precision layout region localization model based on RT-DETR-H trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<td>A highly efficient layout region localization model based on the lightweight PicoDet-S model trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<td>An efficiency-accuracy balanced layout region localization model based on PicoDet-L trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<td>A high precision layout region localization model based on RT-DETR-H trained on a self-built dataset including Chinese and English papers, magazines, and research reports</td>
<tdrowspan="2">PP-OCRv5_rec is a new generation text recognition model. This model aims to efficiently and accurately support the recognition of four major languages: Simplified Chinese, Traditional Chinese, English, and Japanese, as well as complex text scenes like handwriting, vertical text, pinyin, and rare characters with a single model. It balances recognition effectiveness, inference speed, and model robustness, providing efficient and accurate technical support for document understanding in various scenarios.</td>
<td>PP-OCRv4_server_rec_doc is trained on a mix of more Chinese document data and PP-OCR training data based on PP-OCRv4_server_rec, enhancing recognition capabilities for some traditional Chinese characters, Japanese, and special characters, supporting over 15,000+ characters. Besides improving document-related text recognition, it also enhances general text recognition capabilities</td>
> ❗ Listed above are the <b>6 core models</b> that are the focus of the text recognition module, which supports a total of <b>20 full models</b>, including multiple multi-language text recognition models, with the complete model list as follows:
<tdrowspan="2">PP-OCRv5_rec is a new generation text recognition model. This model aims to efficiently and accurately support the recognition of four major languages: Simplified Chinese, Traditional Chinese, English, and Japanese, as well as complex text scenes like handwriting, vertical text, pinyin, and rare characters with a single model. It balances recognition effectiveness, inference speed, and model robustness, providing efficient and accurate technical support for document understanding in various scenarios.</td>
<td>PP-OCRv4_server_rec_doc is trained on a mix of more Chinese document data and PP-OCR training data based on PP-OCRv4_server_rec, enhancing recognition capabilities for some traditional Chinese characters, Japanese, and special characters, supporting over 15,000+ characters. Besides improving document-related text recognition, it also enhances general text recognition capabilities</td>
SVTRv2 is a server-side text recognition model developed by the OpenOCR team from Fudan University's Visual and Learning Lab (FVL), which won first place in the PaddleOCR Algorithm Model Challenge - Task 1: OCR End-to-End Recognition, improving the end-to-end recognition accuracy on the A leaderboard by 6% compared to PP-OCRv4.
<tdrowspan="1">RepSVTR text recognition model is a mobile-side text recognition model based on SVTRv2, which won first place in the PaddleOCR Algorithm Model Challenge - Task 1: OCR End-to-End Recognition, improving the end-to-end recognition accuracy on the B leaderboard by 2.5% compared to PP-OCRv4, with comparable inference speed.</td>
<td>An ultra-lightweight Traditional Chinese recognition model trained based on the PP-OCRv3 recognition model, supporting Traditional Chinese and number recognition</td>
<td>An ultra-lightweight Arabic letter recognition model trained based on the PP-OCRv3 recognition model, supporting Arabic letters and number recognition</td>
<td>An ultra-lightweight Cyrillic letter recognition model trained based on the PP-OCRv3 recognition model, supporting Cyrillic letters and number recognition</td>
<td>An ultra-lightweight Devanagari letter recognition model trained based on the PP-OCRv3 recognition model, supporting Devanagari letters and number recognition</td>
<li>Layout Region Detection Model: PaddleOCR self-built layout region detection dataset, containing 500 common document type images such as Chinese and English papers, magazines, contracts, books, exam papers, and research reports.</li>
<li>3-Class Layout Detection Model: PaddleOCR self-built layout region detection dataset, containing 1154 common document type images such as Chinese and English papers, magazines, and research reports.</li>
<li>17-Class Region Detection Model: PaddleOCR self-built layout region detection dataset, containing 892 common document type images such as Chinese and English papers, magazines, and research reports.</li>
<li>Text Detection Model: PaddleOCR self-built Chinese dataset covering multiple scenarios such as street scenes, web images, documents, and handwriting, where detection includes 500 images.</li>
<li>Chinese Recognition Model: PaddleOCR self-built Chinese dataset covering multiple scenarios such as street scenes, web images, documents, and handwriting, where text recognition includes 11,000 images.</li>
<li>ch_SVTRv2_rec: <ahref="https://aistudio.baidu.com/competition/detail/1131/0/introduction">PaddleOCR Algorithm Model Challenge - Task 1: OCR End-to-End Recognition</a> A leaderboard evaluation set.</li>
<li>ch_RepSVTR_rec: <ahref="https://aistudio.baidu.com/competition/detail/1131/0/introduction">PaddleOCR Algorithm Model Challenge - Task 1: OCR End-to-End Recognition</a> B leaderboard evaluation set.</li>
<li>English Recognition Model: Self-built internal English dataset.</li>
<li>Text Line Orientation Classification Model: Self-built internal dataset covering multiple scenarios such as documents and certificates, containing 1000 images.</li>
<li>Seal Text Detection Model: Self-built internal dataset containing 500 circular seal images.</li>
<b>If you are more concerned with model accuracy, please choose a model with higher accuracy. If you are more concerned with inference speed, please choose a model with faster inference speed. If you are more concerned with model storage size, please choose a model with smaller storage size</b>.
Before using the seal text recognition pipeline locally, please ensure that you have completed the installation of the wheel package according to the [installation tutorial](../installation.md). Once the installation is complete, you can experience it locally via the command line or integrate it with Python.
<details><summary><b>The command line supports more parameter settings. Click to expand for detailed explanations of command line parameters.</b></summary>
Local path of image or PDF file, e.g., <code>/root/data/img.jpg</code>; <b>URL link</b>, e.g., network URL of image or PDF file: <ahref="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/seal_text_det.png">Example</a>; <b>Local directory</b>, the directory should contain images to be predicted, e.g., local path: <code>/root/data/</code> (currently does not support prediction of PDF files in directories; PDF files must be specified with a specific file path).
Whether to load and use the layout detection module. If not set, the parameter will default to the value initialized in the pipeline, which is <code>True</code>.</td>
Such as 0.2, indicates filtering out all bounding boxes with a confidence score less than 0.2. If not set, the default PaddleX official model configuration will be used.
<td>Whether to load and use NMS (Non-Maximum Suppression) post-processing for layout region detection to filter out overlapping boxes. If not set, the default configuration of the official model will be used.</td>
A positive float number, e.g., 1.1, indicating that the center of the bounding box remains unchanged while the width and height are both scaled up by a factor of 1.1. If not set, the default PaddleX official model configuration will be used.
<li><b>large</b>: When set to "large", only the largest outer bounding box will be retained for overlapping bounding boxes, and the inner overlapping boxes will be removed;</li>
<li><b>small</b>: When set to "small", only the smallest inner bounding boxes will be retained for overlapping bounding boxes, and the outer overlapping boxes will be removed;</li>
<li><b>union</b>: No filtering of bounding boxes will be performed, and both inner and outer boxes will be retained;</li>
</ul>If not set, the default PaddleX official model configuration will be used.
<td>Limit type for image side in seal text detection.
Supports <code>min</code> and <code>max</code>; <code>min</code> ensures shortest side ≥ <code>det_limit_side_len</code>, <code>max</code> ensures longest side ≤ <code>limit_side_len</code>. If not set, default is <code>min</code>.
</ul>If not set, the pipeline initialized value for this parameter will be used. During initialization, the local GPU device 0 will be preferred; if unavailable, the CPU device will be used.
* The above command line is for quickly experiencing and viewing the effect. Generally, in a project, you often need to integrate through code. You can complete the quick inference of the pipeline with just a few lines of code. The inference code is as follows:
<td>Whether to load and use the layout detection module. If set to <code>None</code>, the parameter will default to the value initialized in the pipeline, which is <code>True</code>.</td>
<li><b>str</b>: Supports <code>min</code> and <code>max</code>. <code>min</code> ensures the shortest side is no less than <code>det_limit_side_len</code>, while <code>max</code> ensures the longest side is no greater than <code>limit_side_len</code>;</li>
<li><b>None</b>: If set to <code>None</code>, the default value is <code>min</code>.</li>
<li><b>CPU</b>: e.g., <code>cpu</code> means using CPU for inference;</li>
<li><b>GPU</b>: e.g., <code>gpu:0</code> means using GPU 0;</li>
<li><b>NPU</b>: e.g., <code>npu:0</code> means using NPU 0;</li>
<li><b>XPU</b>: e.g., <code>xpu:0</code> means using XPU 0;</li>
<li><b>MLU</b>: e.g., <code>mlu:0</code> means using MLU 0;</li>
<li><b>DCU</b>: e.g., <code>dcu:0</code> means using DCU 0;</li>
<li><b>None</b>: If set to <code>None</code>, the pipeline initialized value for this parameter will be used. During initialization, the local GPU device 0 will be preferred; if unavailable, the CPU device will be used.</li>
<td>Whether to enable high-performance inference.</td>
<td><code>bool</code></td>
<td><code>False</code></td>
</tr>
<tr>
<td><code>use_tensorrt</code></td>
<td>Whether to use TensorRT for accelerated inference.</td>
<td><code>bool</code></td>
<td><code>False</code></td>
</tr>
<tr>
<td><code>min_subgraph_size</code></td>
<td>Minimum subgraph size used to optimize model subgraph computation.</td>
<td><code>int</code></td>
<td><code>3</code></td>
</tr>
<tr>
<td><code>precision</code></td>
<td>Computation precision, e.g., fp32, fp16.</td>
<td><code>str</code></td>
<td><code>"fp32"</code></td>
</tr>
<tr>
<td><code>enable_mkldnn</code></td>
<td>Whether to enable MKL-DNN acceleration. If set to <code>None</code>, MKL-DNN is enabled by default.</td>
<td><code>bool</code></td>
<td><code>None</code></td>
</tr>
<tr>
<td><code>cpu_threads</code></td>
<td>Number of threads used for inference on CPU.</td>
<td><code>int</code></td>
<td><code>8</code></td>
</tr>
<tr>
<td><code>paddlex_config</code></td>
<td>Path to the PaddleX pipeline configuration file.</td>
<td><code>str</code></td>
<td><code>None</code></td>
</tr>
</tbody>
</table>
(2) Call the `predict()` method of the Seal Text Recognition pipeline object for inference prediction. This method will return a `generator`. Below are the parameters and their descriptions for the `predict()` method:
<table>
<thead>
<tr>
<th>Parameter</th>
<th>Parameter Description</th>
<th>Parameter Type</th>
<th>Default Value</th>
</tr>
</thead>
<tr>
<td><code>input</code></td>
<td>Input data to be predicted. Required. Supports multiple types:
<li><b>Python Var</b>: Image data represented by <code>numpy.ndarray</code>;</li>
<li><b>str</b>: Local path of an image or PDF file, e.g., <code>/root/data/img.jpg</code>; <b>URL link</b>, e.g., the network URL of an image or PDF file: <ahref="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/seal_text_det.png">Example</a>; <b>Local directory</b>, containing images to be predicted, e.g., <code>/root/data/</code> (currently does not support prediction of PDF files in directories; PDF files must be specified with an exact file path);</li>
<li><b>List</b>: Elements of the list must be of the above types, e.g., <code>[numpy.ndarray, numpy.ndarray]</code>, <code>[\"/root/data/img1.jpg\", \"/root/data/img2.jpg\"]</code>, <code>[\"/root/data1\", \"/root/data2\"]</code>.</li>
(3) Process the prediction results. The prediction result for each sample is of `dict` type and supports operations such as printing, saving as an image, and saving as a `json` file:
<td>Specify the indentation level to beautify the output <code>JSON</code> data for better readability, effective only when <code>format_json</code> is <code>True</code>.</td>
<td>Control whether to escape non-<code>ASCII</code> characters to <code>Unicode</code>. When set to <code>True</code>, all non-<code>ASCII</code> characters will be escaped; <code>False</code> will retain the original characters, effective only when <code>format_json</code> is <code>True</code>.</td>
<td>Specify the indentation level to beautify the output <code>JSON</code> data for better readability, effective only when <code>format_json</code> is <code>True</code>.</td>
<td>Control whether to escape non-<code>ASCII</code> characters to <code>Unicode</code>. When set to <code>True</code>, all non-<code>ASCII</code> characters will be escaped; <code>False</code> will retain the original characters, effective only when <code>format_json</code> is <code>True</code>.</td>
- Calling the `print()` method will print the results to the terminal, and the explanations of the printed content are as follows:
-`input_path`: `(str)` The input path of the image to be predicted.
-`model_settings`: `(Dict[str, bool])` The model parameters required for pipeline configuration.
-`use_doc_preprocessor`: `(bool)` Controls whether to enable the document preprocessing sub-pipeline.
-`use_layout_detection`: `(bool)` Controls whether to enable the layout detection sub-module.
-`layout_det_res`: `(Dict[str, Union[List[numpy.ndarray], List[float]]])` The output result of the layout detection sub-module. Only exists when `use_layout_detection=True`.
-`input_path`: `(Union[str, None])` The image path accepted by the layout detection module. Saved as `None` when the input is a `numpy.ndarray`.
-`page_index`: `(Union[int, None])` Indicates the current page number of the PDF if the input is a PDF file; otherwise, it is `None`.
-`boxes`: `(List[Dict])` A list of detected layout seal regions, with each element containing the following fields:
-`cls_id`: `(int)` The class ID of the detected seal region.
-`score`: `(float)` The confidence score of the detected region.
-`coordinate`: `(List[float])` The coordinates of the four corners of the detection box, in the order of x1, y1, x2, y2, representing the x-coordinate of the top-left corner, the y-coordinate of the top-left corner, the x-coordinate of the bottom-right corner, and the y-coordinate of the bottom-right corner.
-`seal_res_list`: `List[Dict]` A list of seal text recognition results, with each element containing the following fields:
-`input_path`: `(Union[str, None])` The image path accepted by the seal text recognition pipeline. Saved as `None` when the input is a `numpy.ndarray`.
-`page_index`: `(Union[int, None])` Indicates the current page number of the PDF if the input is a PDF file; otherwise, it is `None`.
-`model_settings`: `(Dict[str, bool])` The model configuration parameters for the seal text recognition pipeline.
-`use_doc_preprocessor`: `(bool)` Controls whether to enable the document preprocessing sub-pipeline.
-`use_textline_orientation`: `(bool)` Controls whether to enable the text line orientation classification sub-module.
-`doc_preprocessor_res`: `(Dict[str, Union[str, Dict[str, bool], int]])` The output result of the document preprocessing sub-pipeline. Only exists when `use_doc_preprocessor=True`.
-`input_path`: `(Union[str, None])` The image path accepted by the document preprocessing sub-pipeline. Saved as `None` when the input is a `numpy.ndarray`.
-`model_settings`: `(Dict)` The model configuration parameters for the preprocessing sub-pipeline.
-`use_doc_orientation_classify`: `(bool)` Controls whether to enable document orientation classification.
-`use_doc_unwarping`: `(bool)` Controls whether to enable document unwarping.
-`angle`: `(int)` The predicted result of document orientation classification. When enabled, it takes values [0, 1, 2, 3], corresponding to [0°, 90°, 180°, 270°]; when disabled, it is -1.
-`dt_polys`: `(List[numpy.ndarray])` A list of polygon boxes for seal text detection. Each detection box is represented by a numpy array of multiple vertex coordinates, with the array shape being (n, 2).
-`dt_scores`: `(List[float])` A list of confidence scores for text detection boxes.
-`text_det_params`: `(Dict[str, Dict[str, int, float]])` Configuration parameters for the text detection module.
-`limit_side_len`: `(int)` The side length limit value during image preprocessing.
-`limit_type`: `(str)` The handling method for side length limits.
-`thresh`: `(float)` The confidence threshold for text pixel classification.
-`box_thresh`: `(float)` The confidence threshold for text detection boxes.
-`unclip_ratio`: `(float)` The expansion ratio for text detection boxes.
-`text_type`: `(str)` The type of seal text detection, currently fixed as "seal".
-`text_rec_score_thresh`: `(float)` The filtering threshold for text recognition results.
-`rec_texts`: `(List[str])` A list of text recognition results, containing only texts with confidence scores above `text_rec_score_thresh`.
-`rec_scores`: `(List[float])` A list of confidence scores for text recognition, filtered by `text_rec_score_thresh`.
-`rec_polys`: `(List[numpy.ndarray])` A list of text detection boxes filtered by confidence score, in the same format as `dt_polys`.
-`rec_boxes`: `(numpy.ndarray)` An array of rectangular bounding boxes for detection boxes; the seal recognition pipeline returns an empty array.
- Calling the `save_to_json()` method will save the above content to the specified `save_path`. If a directory is specified, the saved path will be `save_path/{your_img_basename}_res.json`. If a file is specified, it will be saved directly to that file. Since JSON files do not support saving numpy arrays, `numpy.array` types will be converted to list format.
- Calling the `save_to_img()` method will save the visualization results to the specified `save_path`. If a directory is specified, the saved path will be `save_path/{your_img_basename}_seal_res_region1.{your_img_extension}`. If a file is specified, it will be saved directly to that file. (The pipeline usually contains multiple result images, so it is not recommended to specify a specific file path directly, as multiple images will be overwritten, and only the last image will be retained.)
* Additionally, you can obtain visualized images with results and prediction results through attributes, as follows:
<table>
<thead>
<tr>
<th>Attribute</th>
<th>Description</th>
</tr>
</thead>
<tr>
<tdrowspan="1"><code>json</code></td>
<tdrowspan="1">Get the prediction results in <code>json</code> format.</td>
</tr>
<tr>
<tdrowspan="2"><code>img</code></td>
<tdrowspan="2">Get the visualization results in <code>dict</code> format.</td>
</tr>
</table>
- The prediction results obtained through the `json` attribute are of dict type, with content consistent with what is saved by calling the `save_to_json()` method.
- The prediction results returned by the `img` attribute are of dict type. The keys are `layout_det_res`, `seal_res_region1`, and `preprocessed_img`, corresponding to three `Image.Image` objects: one for visualizing layout detection, one for visualizing seal text recognition results, and one for visualizing image preprocessing. If the image preprocessing sub-module is not used, `preprocessed_img` will not be included in the dictionary. If the layout region detection module is not used, `layout_det_res` will not be included.
## 3. Development Integration/Deployment
If the pipeline meets your requirements for inference speed and accuracy, you can proceed directly with development integration/deployment.
If you need to integrate the pipeline into your Python project, you can refer to the example code in [2.2 Python Script Method](#22-python脚本方式集成).
🚀 High-Performance Inference: In real-world production environments, many applications have stringent performance requirements for deployment strategies, especially in terms of response speed, to ensure efficient system operation and a smooth user experience. To address this, PaddleOCR offers high-performance inference capabilities aimed at deeply optimizing the performance of model inference and pre/post-processing, thereby significantly accelerating the end-to-end process. For detailed high-performance inference procedures, please refer to [High-Performance Inference](../deployment/high_performance_inference.md).
☁️ Service Deployment: Service deployment is a common form of deployment in real-world production environments. By encapsulating inference functionality into a service, clients can access these services via network requests to obtain inference results. For detailed production service deployment procedures, please refer to [Serving](../deployment/serving.md).
Below are the API references for basic serving deployment and multi-language service invocation examples:
<details><summary>API Reference</summary>
<p>For the main operations provided by the service:</p>
<ul>
<li>The HTTP request method is POST.</li>
<li>The request body and response body are both JSON data (JSON objects).</li>
<li>When the request is processed successfully, the response status code is <code>200</code>, and the attributes of the response body are as follows:</li>
</ul>
<table>
<thead>
<tr>
<th>Name</th>
<th>Type</th>
<th>Description</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>logId</code></td>
<td><code>string</code></td>
<td>The UUID of the request.</td>
</tr>
<tr>
<td><code>errorCode</code></td>
<td><code>integer</code></td>
<td>Error code. Fixed as <code>0</code>.</td>
</tr>
<tr>
<td><code>errorMsg</code></td>
<td><code>string</code></td>
<td>Error message. Fixed as <code>"Success"</code>.</td>
</tr>
<tr>
<td><code>result</code></td>
<td><code>object</code></td>
<td>The result of the operation.</td>
</tr>
</tbody>
</table>
<ul>
<li>When the request is not processed successfully, the attributes of the response body are as follows:</li>
</ul>
<table>
<thead>
<tr>
<th>Name</th>
<th>Type</th>
<th>Description</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>logId</code></td>
<td><code>string</code></td>
<td>The UUID of the request.</td>
</tr>
<tr>
<td><code>errorCode</code></td>
<td><code>integer</code></td>
<td>Error code. Same as the response status code.</td>
</tr>
<tr>
<td><code>errorMsg</code></td>
<td><code>string</code></td>
<td>Error message.</td>
</tr>
</tbody>
</table>
<p>The main operations provided by the service are as follows:</p>
<ul>
<li><b><code>infer</code></b></li>
</ul>
<p>Obtain the seal text recognition result.</p>
<p><code>POST /seal-recognition</code></p>
<ul>
<li>The attributes of the request body are as follows:</li>
</ul>
<table>
<thead>
<tr>
<th>Name</th>
<th>Type</th>
<th>Description</th>
<th>Required</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>file</code></td>
<td><code>string</code></td>
<td>The URL of an image or PDF file accessible by the server, or the Base64-encoded content of the file. By default, for PDF files exceeding 10 pages, only the content of the first 10 pages will be processed.<br/>
To remove the page limit, please add the following configuration to the pipeline configuration file:
<pre><code>Serving:
extra:
max_num_input_imgs: null
</code></pre></td>
<td>Yes</td>
</tr>
<tr>
<td><code>fileType</code></td>
<td><code>integer</code> | <code>null</code></td>
<td>The type of file. <code>0</code> indicates a PDF file, <code>1</code> indicates an image file. If this attribute is not present in the request body, the file type will be inferred from the URL.</td>
<td>No</td>
</tr>
<tr>
<td><code>useDocOrientationClassify</code></td>
<td><code>boolean</code> | <code>null</code></td>
<td>Please refer to the description of the <code>use_doc_orientation_classify</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>useDocUnwarping</code></td>
<td><code>boolean</code> | <code>null</code></td>
<td>Please refer to the description of the <code>use_doc_unwarping</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>useLayoutDetection</code></td>
<td><code>boolean</code> | <code>null</code></td>
<td>Please refer to the description of the <code>use_layout_detection</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>layoutThreshold</code></td>
<td><code>number</code> | <code>null</code></td>
<td>Please refer to the description of the <code>layout_threshold</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>layoutNms</code></td>
<td><code>boolean</code> | <code>null</code></td>
<td>Please refer to the description of the <code>layout_nms</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>Please refer to the description of the <code>layout_unclip_ratio</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>layoutMergeBboxesMode</code></td>
<td><code>string</code> | <code>null</code></td>
<td>Please refer to the description of the <code>layout_merge_bboxes_mode</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealDetLimitSideLen</code></td>
<td><code>integer</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_det_limit_side_len</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealDetLimitType</code></td>
<td><code>string</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_det_limit_type</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealDetThresh</code></td>
<td><code>number</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_det_thresh</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealDetBoxThresh</code></td>
<td><code>number</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_det_box_thresh</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealDetUnclipRatio</code></td>
<td><code>number</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_det_unclip_ratio</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
<tr>
<td><code>sealRecScoreThresh</code></td>
<td><code>number</code> | <code>null</code></td>
<td>Please refer to the description of the <code>seal_rec_score_thresh</code> parameter of the pipeline object's <code>predict</code> method.</td>
<td>No</td>
</tr>
</tbody>
</table>
<ul>
<li>When the request is processed successfully, the <code>result</code> in the response body has the following properties:</li>
</ul>
<table>
<thead>
<tr>
<th>Name</th>
<th>Type</th>
<th>Meaning</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>sealRecResults</code></td>
<td><code>object</code></td>
<td>The seal text recognition result. The array length is 1 (for image input) or the actual number of document pages processed (for PDF input). For PDF input, each element in the array represents the result of each page actually processed in the PDF file.</td>
</tr>
<tr>
<td><code>dataInfo</code></td>
<td><code>object</code></td>
<td>Information about the input data.</td>
</tr>
</tbody>
</table>
<p>Each element in <code>sealRecResults</code> is an <code>object</code> with the following properties:</p>
<table>
<thead>
<tr>
<th>Name</th>
<th>Type</th>
<th>Meaning</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>prunedResult</code></td>
<td><code>object</code></td>
<td>A simplified version of the <code>res</code> field in the JSON representation generated by the <code>predict</code> method of the production object, where the <code>input_path</code> and the <code>page_index</code> fields are removed.</td>
</tr>
<tr>
<td><code>outputImages</code></td>
<td><code>object</code> | <code>null</code></td>
<td>See the description of the <code>img</code> attribute of the result of the pipeline prediction. The images are in JPEG format and encoded in Base64.</td>
</tr>
<tr>
<td><code>inputImage</code></td>
<td><code>string</code> | <code>null</code></td>
<td>The input image. The image is in JPEG format and encoded in Base64.</td>
</tr>
</tbody>
</table></details>
<details><summary>Multi-language Service Invocation Example</summary>
for i, res in enumerate(result["sealRecResults"]):
print(res["prunedResult"])
for img_name, img in res["outputImages"].items():
img_path = f"{img_name}_{i}.jpg"
with open(img_path, "wb") as f:
f.write(base64.b64decode(img))
print(f"Output image saved at {img_path}")
</code></pre></details>
</details>
<br/>
## 4. Custom Development
If the default model weights provided by the seal text recognition pipeline do not meet your requirements in terms of accuracy or speed, you can try to <b>fine-tune</b> the existing models using <b>your own domain-specific or application data</b> to improve the recognition performance of the seal text recognition pipeline in your scenario.
Since the seal text recognition pipeline consists of several modules, if the pipeline's performance does not meet expectations, the issue may arise from any one of these modules. You can analyze images with poor recognition results to identify which module is problematic and refer to the corresponding fine-tuning tutorial links in the table below for model fine-tuning.
<table>
<thead>
<tr>
<th>Scenario</th>
<th>Fine-Tuning Module</th>
<th>Fine-Tuning Reference Link</th>
</tr>
</thead>
<tbody>
<tr>
<td>Inaccurate or missing seal position detection</td>
After you complete the fine-tuning training with a private dataset, you can obtain the local model weight files. You can then use the fine-tuned model weights by specifying the local model save path through parameters or by using a custom pipeline configuration file.
#### 4.2.1 Specify Local Model Path via Parameters
When initializing the pipeline object, specify the local model path through parameters. Taking the usage of fine-tuned weights for a text detection model as an example, the demonstration is as follows:
Command line method:
```bash
# Specify the local model path through --doc_orientation_classify_model_dir
# By default, the PP-LCNet_x1_0_doc_ori model is used as the default text detection model. If the fine-tuned model is not this one, modify the model name with --text_detection_model_name
# By default, the PP-LCNet_x1_0_doc_ori model is used as the default text detection model. If the fine-tuned model is not this one, modify the model name with doc_orientation_classify_model_name
#### 4.2.2 Specify Local Model Path via Configuration File
1. Obtain pipeline Configuration File
You can call the `export_paddlex_config_to_yaml` method of the general OCR pipeline object in PaddleOCR to export the current pipeline configuration to a YAML file:
After obtaining the default pipeline configuration file, replace the local path of the fine-tuned model weights in the corresponding position of the pipeline configuration file. For example:
```yaml
......
SubPipelines:
DocPreprocessor:
SubModules:
DocOrientationClassify:
model_dir: null # Replace with the path of the fine-tuned document orientation classification model weights
model_name: PP-LCNet_x1_0_doc_ori # If the name of the fine-tuned model is different from the default model name, please also modify here
module_name: doc_text_orientation
DocUnwarping:
model_dir: null # Replace with the path of the fine-tuned document unwarping model weights
model_name: UVDoc # If the name of the fine-tuned model is different from the default model name, please also modify here
module_name: image_unwarping
pipeline_name: doc_preprocessor
use_doc_orientation_classify: true
use_doc_unwarping: true
......
```
The pipeline configuration file not only contains the parameters supported by the SealRecognition CLI and Python API but also allows for more advanced configurations. Detailed information can be found in the [PaddleX Model pipeline Usage Overview](https://paddlepaddle.github.io/PaddleX/3.0/en/pipeline_usage/pipeline_develop_guide.html), where you can find the corresponding pipeline usage tutorial and adjust various configurations as needed.
3. Load pipeline Configuration File in CLI
After modifying the configuration file, specify the path of the modified pipeline configuration file using the --paddlex_config parameter in the command line. PaddleOCR will read its contents as the pipeline configuration. Example:
When initializing the pipeline object, you can pass the PaddleX pipeline configuration file path or configuration dictionary through the paddlex_config parameter. PaddleOCR will read its contents as the pipeline configuration. Example:
