We have created an original AI-OCR "Flax Scanner for Trade Documents" that automatically recognizes 54 items and extracts text for Commercial Invoices, which are widely used around the world but have different formats for each country and company. "Commercial Invoice general-purpose model" will be on sale from August 9, 2023.

Demo video URL: https://www.youtube.com/watch?v=xR-wPl661Dw

～Important item reading accuracy 91%* / Automatically reads 54 invoice items in different formats～

The development of an AI-OCR system that can read documents with countless formats, such as commercial invoices, requires advanced AI-OCR technology (feature learning type*) that does not require the coordinate definition of extraction items.¹) is required.
The feature-learning AI-OCR "Flax Scanner for Trade Documents/Commercial Invoice General-purpose Model" to be announced this time can automatically read 54 items (*see attached document) organized by frequency of use, priority, etc. , among which 24 items with particularly high importance have an average character accuracy of 91.24%*², average accuracy per item 82.74%*²This enables highly accurate reading.
Cinnamon AI will promote DX in the trading industry with its versatile and highly accurate AI-OCR, which can be introduced quickly without the cost and time required to build a dedicated model.
———-
*1 Feature learning AI-OCR identifies and describes items by learning various common features for each item in multiple forms.This is a technology that allows you to link content and items. Since it automatically determines what type of document it is and what is written where, the customer canNo settings required. Therefore, it can be said that AI-OCR is strong in handling non-standard forms in a wide variety of formats.
*2 All values are Cinnamon AI test data values.

■Development background
One of the challenges in domestic trade operations is the need for people to check and enter required items from trade-related forms (trade documents). In addition, the format of the forms differs depending on the country and company, so experience and skills are required to confirm the necessary information, and input work takes time.

Until now, Cinnamon AI has developed a dedicated AI-OCR that specializes in forms in the trade industry.Flax Scanner for trade documents”, and the “Join the trade consortiumWe have focused on promoting DX by introducing AI in the trading industry, and have built up a track record of implementation and know-how.

The development of a customer-specific AI-OCR model is based on item-specific accuracy (an indicator of whether all items can be read accurately) of 85 to 90%, and character-by-character accuracy (an indicator of how many out of many characters can be read) of 90 to 95%. However, it was costly to develop a dedicated AI-OCR model for each form, and the number of companies that could implement it was limited. In order to introduce AI-OCR to more companies, we will first start developing a "Commercial Invoice general-purpose model" that has high customer needs, and will use Cinnamon AI's advanced AI technology to create a practical model that can read a wide range of Commercial Invoice formats. A general-purpose AI-OCR of this level has been completed and is now being announced.

■High-precision general-purpose AI-OCR that supports various formats realized with advanced AI technology
"Flax Scanner for Trade Documents Commercial Invoice general-purpose model" has been adopted by a wide range of companies that handle trade documents, including exporters/importers (manufacturers/trading companies, etc.), shipping companies, airlines, forwarders, insurance companies (non-life insurance), banks, etc. This was made possible by Cinnamon AI's advanced AI technology.

Most AI-OCR systems are developed using a "coordinate definition type" that predefines where and what information should be written from a formatted document. However, in global trade transactions, there are countless formats, so it is virtually impossible to use the "coordinate definition type".

The "Flax Scanner for Trade Documents Commercial Invoice general-purpose model" announced this time was developed using "feature learning AI-OCR". "Feature Learning AI-OCR" is a learning method that allows AI to learn from a huge amount of data, finds features (patterns and consistency), and then allows the AI to identify what kind of data that data is. . Since there is no need to define coordinates in advance, it is possible to handle trade documents with an infinite number of formats. Advanced AI technology is required to support reading a wide range of formats, and Cinnamon AI excels in this technical area called IDP (intelligent document processing = automatic knowledge extraction technology).

Cinnamon AI will develop and expand the introduction of general-purpose AI-OCR for each form, such as Packing List, Bill of Lading, and Arrival Notice, which are in need following Commercial Invoice. We aim to

■Main companies that can be introduced
Exporters/importers (manufacturers/trading companies, etc.), shipping companies, airlines, forwarders, insurance companies (non-life insurance), banks

————————————————————————————————————–

\We will be holding a seminar introducing "Flax Scanner for Trade Documents Commercial Invoice General-purpose Model"./

■Date and time: August 23rd (Wednesday) 12:00-12:40
■Participation application URL:https://go.cinnamon.ai/20230823_LP1.html

moreover,International Logistics Exhibition 2023(September 13th to September 15th/Tokyo Big Sight) We will be demonstrating the "Flax Scanner for Trade Documents Commercial Invoice General-purpose Model"! Please stop by our booth.
https://www.logis-tech-tokyo.gr.jp/ie/index.html

————————————————————————————————————–

◇ Contact information for inquiries from companies regarding this matter(Cinnamon AI inquiry form): https://go.cinnamon.ai/inquiry.html
*Introducing the "Flax Scanner for Trade Documents Commercial Invoice general-purpose model" allows for system use using the UI, providing only the API, and customization proposals for each company.

<Reference materials>
 "Flax Scanner for trade documents Commercial Invoice general purpose model"
54 items to read

Based on interviews with a wide range of companies that handle trade documents, including exporters/importers (manufacturers/trading companies, etc.), shipping companies, airlines, forwarders, insurance companies (non-life insurance), and banks, we have developed the following 54 items as a commercial invoice general-purpose model. The item has been decided to be read. (Highly important items are *)
Items marked with (details) are read in table format.

◇Reading accuracy (test data value by Cinnamon AI)
Important item accuracy ●Character unit accuracy: average 91.24% / ●item accuracy: average 82.74%
All item accuracy ●Character unit accuracy: average 87.23% / ●item accuracy: average 78.86%

No.given name	Reading items
1	Shipper/Exporter Name*
2	Shipper/Exporter Address*
3	Shipper/Exporter Phone No.
4	Shipper/Exporter FAX No.
5	Shipper/Exporter Department
6	Consignee/Messrs/Shipped to Name*
7	Consignee/Messrs/Shipped to Address*
8	Consignee/Messrs/Shipped to Phone No.
9	Buyer/Accountee/Sold to Name*
10	Buyer/Accountee/Sold to Address*
11	Buyer/Accountee/Sold to Phone No.
12	Invoice Date
13	Invoice No. *
14	B/L No.
15	Customer No./Account No.
16	Contract No./Contract No.
17	Order No.
18	P/O No.
19	Booking No.
20	Vessel Name/Means of Transport & Route*
21	Voyage No. *
22	Port of Loading/From*
23	Shipment Date/Date of Departure/ON
24	Via
25	Port of Discharge*
26	Final Destination
27	Country of Origin*
28	Payment Terms*
29	L/C No.
30	Item Name/Description of Goods*
31	HS Code (details)*
32	Item No./Lot No. (details)*
33	Item Quantity/Unit Quantity*
34	Total Quantity per item*
35	Unit*
36	Unit Price (details)*
37	Amount per item*
38	Tax*
39	Case Mark/Marks & Nos./No. of Packages
40	Incoterms/Trade Terms*
41	Total Quantity/Total Packages
42	Subtotal Amount
43	Freight Cost
44	Bank Name
45	Bank Branch Name
46	Account Name
47	Bank Account Number
48	Bank Address
49	Swift Code/Swift Address/BIC Code
50	Payment Date/Due Date
51	Signatory Company
52	Signed by
53	Currency*
54	IBAN Code