Handwritten tax receipt recognition method based on improved DBNet and CRNN
Lan Wei, Jingqi Sun · 2026
Read Handwritten tax receipt recognition method based on improved DBNet and CRNN by Lan Wei, Jingqi Sun (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0352963.txt.
Description
Lan Wei and Jingqi Sun present this 2026 PLOS One computer vision paper on reading handwritten tax receipts automatically. Tax offices handle mountains of messy handwriting, and the authors pair an improved text-spotting network with an improved reading network to find and transcribe receipt text. Their detection stage adds a feature-enhancement module to sharpen its view of faint handwriting.
In testing, the detector located text with high precision and speed, beating two recent rival systems on both counts. The reader likewise transcribed text more accurately and quickly than the comparison methods. The authors present the combined system as a practical way to cut the manual workload of invoice processing while keeping errors low.
Catalog record
| CREATOR | Lan Wei, Jingqi Sun |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44726 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |