Training during invoice capture
If you are not satisfied with the quality of data that ABBYY FlexiCapture for Invoices automatically extracts from fields, you can train the program while capturing invoices.
This is done by indicating the correct location of fields on the three images of this type of invoice. After the export all the documents are moved to the Training stage by default. It will perform the training procedure and use the results of this procedure when processing new invoices of this type.
Important! Training during invoice capture must be enabled by the Administrator.
How to train the program while capturing invoices
1. Collect a batch of invoices (e.g. invoices received during the past month) and begin inputting them ( See How to capture invoices).
2. After input, documents are automatically recognized (if this option is enabled by the Administrator) and checked using rules.
3. If the resulting status of an invoice is other than Valid or you have other reasons to believe that the program failed to find fields on the image or located them incorrectly, open the invoice in the document editor.
4. Check the document form. The Vendor group of fields must be filled out correctly.
Training is performed separately on each type of document. Invoices from a single vendor are considered to belong to a single type. If the vendor on an invoice is determined incorrectly, that invoice will be assigned to the wrong document type during training. If the vendor was not determined, select a vendor from the database. If you cannot find the right vendor in the database, type the data from the image in the fields from the Vendor group and save these changes to the database by clicking the Save Changes button.
Depending on your project's settings, you may also have to specify the unique ID of an invoice's vendor in order to enable the program to train on that invoice. To do this, type the unique ID in the VATID field (this field may have a different name in some localizations of projects). The VATID is a unique identification number assigned to companies for value-added taxation purposes.
5. One of the necessary conditions for training is that the positions of the regions of all fields on the images match the real positions of the fields on the invoice. You have to make sure that the regions are located correctly.
Correct the positions of the regions on images of at least three documents and draw regions for fields that were not detected.
How to change the region of a field
- Place the mouse pointer in a desired field on the data form, find the corresponding region on the image (it will be highlighted in blue), and click it (or draw a rectangle with the mouse).
- Position the mouse pointer over a desired region on the image (it will be highlighted in blue), click it (or draw the region with the mouse), and then select the corresponding field from the drop-down list that opens.
- Adjust the position of a region on the image by moving its boundaries with the mouse.
- Delete an incorrectly located region from the image: position the mouse pointer over its rectangle and when a red cross appears in the top right corner, click the red cross. The markup of the region will be deleted. Now create a new region for this field in the right location.
- On the data form, start typing a value into a field. A drop-down list will be displayed listing the words captured from the image that resemble the word that you are typing. Select the right word from the list, and the position of the word on the image will become the region of the field.
6. After the export all the documents are moved to the Training stage by default. The training process will be started automatically (the Train Document Definition window will be displayed). Either a special FlexiLayout Variant will be created for this document type or the program will prompt you to continue adding sample images (if this is the case, select the next document and go back to step 4).
If the FlexiLayout Variant is successfully trained, it will be used for the next invoice from the vendor whose invoices belong to the given document type. Upon recognition, the regions of the fields will be matched against the fields on invoice image, taking into account the knowledge obtained through training.
7. Add a few more invoices from the vendor whose document type has been trained and recognize them. Then open the newly added invoices one by one in the document editor to check if the regions are marked up correctly. If all the regions are located correctly, no additional training is required.
If you are not satisfied with the results achieved by completing the steps in this section, continue training on invoices from the vendor (steps 4-5).