16.3.1
📅 2026-03-27
New Features
CDocumentProcessing
This version introduces the CDocumentProcessing class.
It is an all-in-one API that handles the entire document-processing workflow.
With a single call, it allows you to convert a series of input images into a set of any supported output formats.
It also manages multithreading for you, using an inter-page multithreading strategy to optimize processing time. Memory consumption is optimized as pages are discarded once they have been processed.
The class provides convenient methods to retrieve the document-processing result:
-
A callback function
-
A
std::futureobject -
An awaiter object (for C++20 and later) for use with
co_await
| This API is experimental. It is fully functional, but it can change in future versions of iDRS. |
Improvements
High Quality OCR (HQOCR) for Chinese and Hebrew
This version adds support for Chinese Simplified, Chinese Traditional, and Hebrew in our High‑Quality OCR (HQOCR) engine. With these additions, HQOCR now covers all remaining languages in its roadmap. This enhancement delivers significant improvements in both accuracy and performance, especially on challenging documents such as low-resolution images, magazines, and newspapers.
Added/removed resources
Binaries removal
This version removes the following binaries from the package. If these files are still in your installation, you can safely delete them.
| Windows | Linux | macOS |
|---|---|---|
DLL files formerly in |
Shared libraries formerly in |
Dynamic libraries formerly in |
idrsdmtxbarcodewrapper16.dll |
libidrsdmtxbarcodewrapper16.so |
libidrsdmtxbarcodewrapper16.dylib |
OCR Resources file changes
Because this version replaces the OCR engine for Chinese and Hebrew, the following resource files are added or removed:
| Resource file | Status |
|---|---|
chinese_generic_rnn_engine.bin |
Added |
net_schinese_rnn.bin |
Added |
net_tchinese_rnn_hori.bin |
Added |
net_tchinese_rnn_vert.bin |
Added |
hebrew_engine.bin |
Added |
net_hebrew.bin |
Added |
chinese.bin |
Removed |
chinese_bigrams.bin |
Removed |
chinese_features.bin |
Removed |
chinese_features_fast.bin |
Removed |
chinese_features_one_candidate.bin |
Removed |
chinese_font_features.bin |
Removed |
chinese_generic.bin |
Removed |
chinese_same_proba.bin |
Removed |
chinese_similar_characters_differentiation.bin |
Removed |
chinese_traditional.bin |
Removed |
chinese_traditional_bigrams.bin |
Removed |
chinese_traditional_features.bin |
Removed |
chinese_traditional_features_fast.bin |
Removed |
chinese_traditional_features_one_candidate.bin |
Removed |
chinese_traditional_font_features.bin |
Removed |
chinese_traditional_same_proba.bin |
Removed |
chinese_traditional_similar_characters_differentiation.bin |
Removed |
hebrew.ocr |
Removed |
Fixed bugs
| Internal ID | Description | Service desk IDs |
|---|---|---|
IDRSRD-5903 |
The DRS engine misses some text when the segmentation filters are switched off |
|
IDRSRD-9867 |
iDRS does not properly detect the table on the attached image |
|
IDRSRD-10063 |
[Regression] Pdf output should allow to put empty-string title, when universal access is disabled" |
|
IDRSRD-10071 |
AnalyzePage() and RecognizeText() hanging when std::fesetround(FE_UPWARD) is set. |
ISD-37489 |
IDRSRD-10087 |
Performing pdf compression without OCR |
ISD-38280 |
IDRSRD-10104 |
The iDRS failed creating a DOCX on a specific image |
|
IDRSRD-10110 |
iDRS setting TextDetectionSensivity.MaximumText does not grab enough test on a customer test image |
|
IDRSRD-10111 |
IDRS_ERROR_CHARACTER_RECOGNITION_ENGINE_FAILURE on attached pdf |
ISD-37695 |
IDRSRD-10120 |
The iDRS crashes when processing the attached document, in traditional Chinese" |
ISD-36631 |
IDRSRD-10122 |
The speed and memory performance of CPageProcessing can be improved |
|
IDRSRD-10123 |
The iDRS should allow better control of its internal threading configuration |
ISD-34219 |
IDRSRD-10124 |
Speed performance of iDRS iHQC feature can be improved |
|
IDRSRD-10129 |
Skew detection fails on small images |
|
IDRSRD-10130 |
Memory consumption of iDRS iHQC feature can be improved |
|
IDRSRD-10132 |
iDRS misses text from a Korean image |
ISD-37746 |
IDRSRD-10148 |
"iDRS can overwrite existing file even if missing permissions, on Linux platform" |
|
IDRSRD-10166 |
iDRS version 16.2.6+178 throws an exception when deskewing near-blank color images |
|
IDRSRD-10172 |
PDFs created from greyscale images by iDRS 16.2.6+178 cannot be opened by Acrobat Reader |
|
IDRSRD-10173 |
"Image skew is not corrected when using iDRS 16.2.6+178 CPageProcessing, if both language and orientation detection are not enabled" |
|
IDRSRD-10175 |
The iDRS fails to run twice Chinese OCR with the same CPageProcessing instance |
ISD-36631 |
IDRSRD-10187 |
Dimensions of generated PDF pages can be more precise |
|
IDRSRD-10199 |
iDRS 16 doesn’t use custom image plugin for file loading |
|
IDRSRD-10211 |
developer guide Aztec 2D barcode example image is actually PDF417 |
|
IDRSRD-10215 |
Regression - CDocument’s page collection is not notified on completion of output processing |
|
IDRSRD-10218 |
iDRS crashes when trying to sign an existing document in MacOS |