Paper invoice information extraction method and device

The embodiment of the invention provides a paper invoice information extraction method and device which are used for improving the paper invoice information extraction speed and accuracy. The method comprises the following steps: acquiring a paper invoice image; performing full face text recognition on the paper invoice image through an optical character recognition (OCR) model to obtain a text recognition result; wherein the OCR model is obtained by combining a target detection model YOLO4 and a text recognition model CRNN, the YOLO4 model is used for carrying out text detection on the paper invoice image, and the CRNN model is used for carrying out text recognition on a text detection result of the YOLO4 model; and exporting text information corresponding to the text recognition result. 本发明实施例提供了一种纸质发票信息提取方法及装置，用于提升纸质发票信息提取速度和准确性。所述方法包括：获取纸质发票图像；通过光学字符识别OCR模型对所述纸质发票图像进行全票面文本识别，得到文本识别结果；其中，所述OCR模型为将目标检测模型YOLO4和文本识别模型CRNN结合得到的，所述YOLO4模型用于对所述纸质发票图像进行文本检测，所述CRNN模型用于对所述YOLO4模型的文本检测结果进行文本识别；导出所述文本识别结果对应的文本信息。