Mobile OCR or Server Based Back End Processing
Usage Scenarios
| Mobile OCR | Server based OCR | |
|---|---|---|
| Application Areas | On Device OCR for instant text re-use e.g. store captured text as notes, apply text to speech, send info via mail or send text to translation service | Server based OCR allows Data extraction of specific fields of interest Usage of big databases (verification & additional meta data) |
| Information to be processed | Printed Text | Printed text, Hand filled survey forms, Checkmarks, Barcodes |
| Application Types | “Thick” Apps with OCR technology included, OS native, | “Thin” Client approach Image capture, data communication with the back-end, More generic Apps for multiple devices for simple task and quality check before the documents are sent to the server |
Technological Differences
| Mobile OCR | Server based OCR | |
|---|---|---|
| CPUs | ARM, x86, Up to 1 GHz, Dual core in new devices (announced 04/2011) | x86, > 3 GHZ Multi core systems, Scalable virtual systems |
| Device Memory | (still) limited | de facto unlimited |
| OCR Memory Requirements | Latin & Cyrillic OCR 8 MB ROM, 10 MB RAM | > 128 MB |
| Storage | limited | unlimited |
| Technology | Compact & tuned to make it work on mobile devices | Broad set of technologies can be used for image preprocessing, document analysis, character recognition, data extraction, |
| Meta Data available for verification | limited, comes with the app | unlimited |
| “Online” Connection | OCR is executed “off-line” | connection required |
| Output Options | Text only Business Card fields | Text, Editable Office Formats, XML - with page layout, Data XML Searchable, compressed PDF ebook Formats Structured Data |
Back to: Comparisons
