In 2021, our stack of technologies steadily grew with each weekly release. Our team focused a lot on OCR, PDF/A generation and validation, and PDF redaction, always with an eye on all our areas of expertise. As a result, in 2021, all our engines became stronger than ever.

Before we go into further detail, here are a few links to check our progress: 

You can also register to the GdPicture.NET & DocuVieware newsletter to get an overview of the updates each month.

• PDF/A-4 generation, conversion, and validation support.

Major improvements of all the PDF engines: parser, generation, rendering, redaction, merging, grayscale conversion, repair, PDF/A conversion & validation. 

Lots of new features in the document conversion domain in 2021:

• HTML format support for viewing and conversion,
• support for HEIF/HEIC format,
• support for EML format,
• support for MSG format,
• support to convert any document to SVG,
• support for file attachments in Office formats.

And as usual, all the other engines also get significant improvements, especially:

• office formats rendering engine,
• generation speed for PNG images,
• JPEG compression in TIFF images,
• RTL support for annotation editing,
• image rendering speed,
• MSG rendering engine,
• SVG rendering engine.

New special contexts:

• New OCRSpecialContext enumeration: MICRLineE13B & MICRLineCMC7,
• New OCRSpecialContext enumeration members: NumericLineML & HandwrittenNumericBoxML. See our blog article Introducing the New GdPicture.NET Deep Learning-based ICR Engine

Major improvements:

• Global improvement of the OCR and text extraction engines,
• Support for Chinese (accuracy, PDF OCR generation),
• MRZ reader accuracy & speed,
• MICR engine accuracy & speed (+50% performance).

New: PDF Reducer engine can now produce PDF/A output.

Improvements:

• Color Detection engine accuracy,
• MRC engine accuracy, speed, & compression rate,
• PDF Reducer engine & compression rate,
• Lossy JBIG2 encoder (better pattern matching on text content).

Just before the end of the year, we released MaxiCode reading and writing support. This barcode symbology is especially useful for building applications in the postal and shipping domains. 

Since we’re working in continuous integration, all the other barcode engines (1D, DataMatrix, PDF417, QR code & micro QR code, Aztec code) have improved accuracy, speed, and robustness.

• New word segmentation support (GdPictureSegmenter),
• Improved automatic page orientation engine speed & accuracy,
• Improved page orientation detection engine (accuracy),
• Improved blank page detection algorithm,
• Improved TWAIN scanning support and support for unstable TWAIN drivers,
• Improved GdViewer Winform rendering engine.

Our SDKs are now available on nuget.org. 

Other general improvements:

• toolkit speed,
• internal rendering engine speed.
• COM edition stream reading performance,
• HTTP transfer support and new support for HTTP/2 protocol.

What’s new?

• Support for 64-bit TWAIN devices acquisition,
• Support for file attachment annotations,
• Chinese traditional locale,
• Contextual menu item to remove selected text.

Improvements:

• UI, UI responsiveness, & UX,
• document loading speed,
• Printing speed,
• HTTP loading speed (up to 4x faster),
• Page rotation mechanism,
• Extended page reordering support to all document types,
• Page transfer & generation speed,
• Page generation speed,
• Form fields support.

In February, we welcomed the UK-based company Aquaforest.

Aquaforest’s solutions are complementary to ours with, among other enterprise software,

• a high performance automated batch OCR and conversion to searchable PDFs for Windows servers (Autobahn DX),
• an audit, OCR conversion, and metadata tagging tool for SharePoint and Office 365 documents (Aquaforest Searchlight),
• a no-code, low code automation tool for the Microsoft platform Power Automate (PDF Connector).

We were thrilled to meet the team in Aylesbury as soon as the sanitary conditions let us, and we enjoyed discovering the Oxford area and touring the famous British pubs (for those of us fans of Inspector Morse and Midsomer Murders, you know what I’m talking about!).

Going through the process of an acquisition in a foreign country during a pandemic (and Brexit!) is an adventure. But we did it, and it was nice to meet in person finally and celebrate this major milestone for our company.

Last October, the PDF Association held a series of webinars about various aspects of the PDF format. Our technical presentations covered PDF/A conversion and validation and the growing low code/no code ecosystem for PDF tools.
We cannot wait to go back to the in-person conference next year. More information on that soon!

We recently bought an extension for our French offices in Muret.

The idea is to have a collaborative space that is more than a workspace. These last two years made everyone rethink what an office should be. Most of us stayed home for the past few months, some of us will not come back to an office ever. But since a significant part of our team wishes to come, why not make their stay a bit more comfortable?
In our new space, the first floor is dedicated to a fully-equipped gym that even includes a boxing ring.
Upstairs, a social space on one side, and on the other side several soundproofed offices with standing desks, and a small kitchen.

We are looking forward to hosting small events there too!

At the beginning of the year, we published an article about our customer INP Toulouse, whose team processes more than one million sheets each year in a very short time frame. Learn more how they leverage our scanning, OCR, barcoding, and image processing technologies to automate the processing of examination papers. 

Along with the release of the first version of our first ICR engine, we wrote a series of article introducing deep learning for OCR:

• Deep learning for OCR
• Deep learning for text detection – part 1
• Deep learning for text detection – part 2
• Deep learning for text recognition

Thank you again for your support in 2021!
2022 is going to bring more exciting things, so stay tuned ?

Cheers,

Elodie