Invoice and receipt recognition
This topic describes the features, benefits, and application scenarios of Alibaba Cloud Optical Character Recognition (OCR) for invoices and receipts. It also provides links to the product APIs.
Product introduction
Based on the deep learning algorithms of Duguang OCR, the invoice and receipt recognition service performs structured recognition of various invoices used for financial reimbursement and tax accounting. The service supports nearly 20 common invoice types, such as value-added tax (VAT) invoices, VAT roll invoices, train tickets, quota invoices, air transport itineraries, taxi receipts, general machine-printed invoices, toll invoices, and bus and ship tickets. The service can also automatically segment and recognize mixed invoices.
Try the feature: https://duguang.aliyun.com/experience?type=bill
Activate to obtain a free quota: https://ocr.console.aliyun.com/overview
Purchase: https://common-buy.aliyun.com/?commodityCode=ocr_invoice_dp_cn#/buy
Product features
Mixed invoice recognition
The mixed invoice recognition feature uses powerful deep learning algorithms and structured template optimization technology from DAMO Academy to ensure industry-leading accuracy for all fields. The service recognizes key fields, such as invoice code, total amount including tax, total amount, purchaser's taxpayer identification number, and issue date, with an accuracy rate of over 99%.
Ride-hailing itinerary recognition
The ride-hailing itinerary recognition feature supports all fields on a ride-hailing itinerary, such as service provider, request date, trip start time, trip end time, passenger's phone number, and total amount, with a recognition accuracy of over 96%.
VAT invoice recognition
The VAT invoice recognition feature supports various invoice types, such as special VAT invoices, general VAT invoices, and electronic VAT invoices. It recognizes key fields, including invoice code, invoice number, issue date, amount, tax amount, verification code, purchaser's taxpayer identification number, seller's taxpayer identification number, and invoice details, with a recognition accuracy of over 99%.
Train ticket recognition
The train ticket recognition feature recognizes fields on train tickets, such as ticket number, departure station, arrival station, departure time, fare, seat class, passenger information, seat number, and train number. The recognition accuracy for key fields, such as name, departure time, departure station, and seat number, is over 96%.
Quota invoice recognition
The quota invoice recognition feature recognizes key fields, including invoice number, invoice code, and amount.
Air transport itinerary recognition
The electronic air transport itinerary recognition feature recognizes fields such as passenger name, ID card number, e-ticket number, date of issue, and issuing entity.
Taxi receipt recognition
The machine-printed taxi receipt recognition feature recognizes key fields, including invoice code, invoice number, date, and amount. The recognition accuracy for key fields, such as invoice code and invoice number, is over 95%.
VAT roll invoice recognition
The VAT roll invoice recognition feature recognizes key fields on roll invoices, including invoice code, invoice number, issue date, amount, verification code, amount in uppercase, seller's taxpayer identification number, and purchaser's taxpayer identification number.
Unified invoice for motor vehicle sales recognition
The unified invoice for motor vehicle sales recognition feature recognizes key fields on vehicle purchase invoices, including invoice code, invoice number, issue date, amount, VAT amount, certificate number, purchaser's name, and purchaser's ID number or code.
Unified invoice for used car sales recognition
The unified invoice for used car sales recognition feature recognizes key fields on used car purchase invoices, including invoice code, invoice number, issue date, amount, purchaser's name, and purchaser's ID number.
General machine-printed invoice recognition
The general machine-printed invoice recognition feature recognizes key fields, including invoice code, invoice number, seller's name, seller's identification number, purchaser's name, purchaser's identification number, and total amount.
Toll invoice recognition
The toll invoice recognition feature recognizes key fields, including invoice code, invoice number, amount, date, vehicle type, exit, and entrance, with an accuracy rate of over 95%.
Bus and ship ticket recognition
The bus and ship ticket recognition feature recognizes key fields, including title, invoice number, departure station, arrival station, date, and amount.
Tax clearance certificate recognition
The tax clearance certificate recognition feature recognizes key fields, including tax authority, taxpayer identification number, taxpayer name, total amount, issuer, and tax payment details, with an accuracy rate of over 98%.
E-commerce order page recognition
The e-commerce order page recognition feature recognizes key fields, including order number, shipping information, monetary amount, store name, and item details.
Payment details page recognition
The payment details page recognition feature recognizes key fields, including payee name, total amount, payment method, product description, and payment time.
Non-tax revenue invoice recognition
The non-tax revenue invoice recognition feature recognizes key fields, including bill code, payer, bill number, total amount, and payee, with an accuracy rate of over 97%.
Hotel folio recognition
The hotel folio recognition feature recognizes key fields, including room number, check-in date, check-out date, total consumption, total payment, and consumption details.
Shopping receipt recognition
The shopping receipt recognition feature recognizes key fields, including issuer name, issue date, phone number, address, and total (actual) amount.
Bank acceptance bill recognition
The bank acceptance bill recognition feature recognizes key fields, including date of issue, bill number, drawer information, payee information, drawee information, and bill amount.
Invoice verification
Duguang OCR supports authenticity verification for five types of VAT invoices: special VAT invoices, general VAT invoices, blockchain invoices (Shenzhen area), unified invoices for motor vehicle sales, and unified invoices for used car sales. This feature does not recognize invoice content. Before verification, you must check the required invoice fields and enter them manually. The invoice issue date must be within the last five years. Invoices can be verified on the day after they are issued.
For more information, see Invoice and receipt verification.
Benefits
High accuracy: The service achieves industry-leading accuracy by training its models on large volumes of image samples.
Low latency: The service provides low latency and elastic scaling using Alibaba Cloud's self-built EAS online service cluster and optimized inference technology.
Cutting-edge technology: The service is built on Alibaba Cloud Platform for AI and uses PAI-TensorFlow, a deep learning framework optimized by Alibaba, to train advanced text detection and recognition models.
Stable service: The service offers elastic scaling based on call volume, provides high extensibility, and ensures service stability during continuous algorithm iterations and optimizations.
Scenarios
Financial reimbursement: Automate the recognition and entry of various tax invoices and transportation tickets. This reduces manual accounting workloads, lowers labor costs, and automates financial reimbursement.
Mobile insurance application: Extract structured information from various insurance documents for use in scenarios such as policy review and premium issuance. This improves the efficiency of information entry, reduces manual accounting, and lowers labor costs.
Financial sharing: Use the service for shared expense reimbursement and expenditure control, shared fund settlement, and shared tax and invoice management. This improves the efficiency of information entry, reduces manual accounting, and lowers labor costs.
Quick access to APIs
Alibaba Cloud Marketplace API access (Legacy) | Official website API access (New) |
Tax clearance certificate recognition | |
Ride-hailing itinerary recognition | |
Toll invoice recognition | |
General machine-printed invoice recognition | |
Bus and ship ticket recognition | |
E-commerce order page recognition | |
Hotel folio recognition | |
Payment details page recognition | |
Non-tax revenue invoice recognition | |
Shopping receipt recognition |