Image character acquisition device and program
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- DAI NIPPON PRINTING CO LTD
- Filing Date
- 2024-12-13
- Publication Date
- 2026-06-25
AI Technical Summary
【0007】 本発明によれば、非定型書類を含む書類の文書画像を入力して項目名と項目値との組み合わせを簡単かつ適切に出力する画像文字取得装置及びプログラムを提供することができる。
Smart Images

Figure 2026104070000001_ABST
Abstract
Claims
1. A document image acquisition method for acquiring document images of documents, The document image acquisition means inputs the document image acquired by the document image acquisition means and an instruction sentence instructing the generation AI to extract items contained in the document image, thereby providing an item combination acquisition means that acquires the combination of item names and item values contained in the document image output by the generation AI. An item combination output means that outputs the combination of the item name and the item value obtained by the item combination acquisition means, An image and character acquisition device equipped with the following features.
2. In the image character acquisition device according to claim 1, The item combination acquisition means is an image character acquisition device that acquires the item name and item value combination output by the generating AI by inputting the document image and an item acquisition instruction statement which instructs the generating AI to extract the item name and item value contained in the document image and then acquire the combination of the item name and item value.
3. In the image character acquisition device according to claim 1, The aforementioned item combination acquisition means is By inputting the document image and an item name acquisition instruction statement that instructs the AI to extract item names from the document image into the generating AI, the item names output by the generating AI are obtained. By inputting the document image and an item value acquisition instruction statement that instructs the AI to extract item values from the document image into the generating AI, the item values output by the generating AI are acquired. Image character acquisition device that acquires the combination of item name and item value output by the generating AI by inputting the document image and an item acquisition instruction statement which includes the acquired item name and item value and instructs the generating AI to acquire a combination of item name and item value based on the document image.
4. In the image character acquisition device according to claim 1, The aforementioned item combination acquisition means is By inputting the document image and an item name acquisition instruction statement that instructs the AI to extract item names and location information from the document image into the generating AI, the AI acquires the item names and location information output by the generating AI. By inputting the document image and an item value acquisition instruction statement that instructs the AI to extract item values and location information from the document image into the generating AI, the AI acquires the item values and location information output by the generating AI. An image character acquisition device that acquires the combination of item name and item value output by the generating AI by inputting an item acquisition instruction statement to the generating AI, which includes the acquired item name and location information, and the acquired item value and location information, and which instructs the generating AI to acquire the combination of item name and item value.
5. In the image character acquisition device according to claim 1, The document image acquisition means acquires an item name-emphasized document image in which the item names in the document image are emphasized, and an item value-emphasized document image in which the item values in the document image are emphasized, The aforementioned item combination acquisition means is By inputting the item name highlighting document image and an item name acquisition instruction statement that instructs the AI to extract item names and location information from the item name highlighting document image into the generating AI, the AI acquires the item names and location information output by the generating AI. By inputting the item value highlighting document image and an item value acquisition instruction statement that instructs the AI to extract item values and location information from the item value highlighting document image into the generating AI, the AI acquires the item values and location information output by the generating AI. An image character acquisition device that acquires the combination of item name and item value output by the generating AI by inputting an item acquisition instruction statement to the generating AI, which includes the acquired item name and location information, and the acquired item value and location information, and which instructs the generating AI to acquire the combination of item name and item value.
6. In the image character acquisition device according to any one of claims 2 to 5, The aforementioned item acquisition instruction statement includes an instruction to infer the item name if the item name is omitted, in an image character acquisition device.
7. In the image character acquisition device according to claim 1, A character region detection means for detecting character regions from the document image acquired by the document image acquisition means, A character recognition process is performed on the character region detected by the character region detection means to obtain a character string; The string acquisition means acquires the item from the string acquired by the string acquisition means, An emphasis processing means for emphasizing the items acquired by the item acquisition means, Equipped with, The item combination acquisition means is an image character acquisition device that acquires combinations of item names and item values output by the generating AI by inputting the document image after enhancement processing by the enhancement processing means and an item acquisition instruction sentence that instructs the generating AI to extract the colored items.
8. In the image character acquisition device according to claim 7, The item acquisition means refers to an item name list storage unit that stores an item name list which is a list of item names, and acquires the strings included in the item name list as the item names, in an image character acquisition device.
9. In the image character acquisition device according to claim 8, The item acquisition means is an image character acquisition device that refers to an item value condition storage unit which stores conditions for item values corresponding to item names, and acquires the string that matches the conditions for the item value as the item value corresponding to the item name.
10. In the image character acquisition device according to any one of claims 7 to 9, The enhancement processing means is an image character acquisition device that enhances the item name and the item value corresponding to the item name by coloring them in the same color.
11. Computers, A document image acquisition method for acquiring document images of documents, The document image acquired by the document image acquisition means and an item acquisition instruction statement that instructs the AI to extract the item names and item values contained in the document image and then acquire the combination of item names and item values are input to the generating AI, thereby providing the item combination acquisition means that acquires the combination of item names and item values contained in the document image output by the generating AI, An item combination output means that outputs the combination of the item name and the item value obtained by the item combination acquisition means, A program to make it work.