Описание
Tesseract is an open source OCR engine. In version 5.5.3 and earlier, prior .traineddata hardening added bounds checks to NetworkIO::CopyTimeStepGeneral and NetworkIO::Randomize in src/lstm/networkio.cpp but left NetworkIO::WriteTimeStepPart and NetworkIO::AddTimeStepPart unchecked. In LSTM::Forward in src/lstm/lstm.cpp, source_ is sized from the independently deserialized na_ field while the WriteTimeStepPart count is ns_, which comes from the CI gate WeightMatrix dim1() value. A crafted NT_LSTM layer can make ns_ much larger than na_, causing a heap out-of-bounds write during the first recognition step on the default LSTM engine and resulting in heap corruption, a crash, or potentially controlled corruption. No fixed release is available as of this review.
A flaw was found in Tesseract, an open-source Optical Character Recognition (OCR) engine. A remote attacker could exploit a heap out-of-bounds write vulnerability in the LSTM::Forward function by providing a specially crafted input. This could lead to heap corruption, application crashes, or potentially controlled memory corruption, which may result in arbitrary code execution or a denial of service.
Отчет
This vulnerability is rated as Important severity because, although triggering it requires loading a manipulated language model file, successful exploitation can result in arbitrary code execution within the context of the executing user or application. In standard Red Hat deployments, Tesseract processes trusted data packages provided by the operating system, which limits exposure unless an environment permits user-supplied models. Systems that integrate the OCR engine into automated ingestion pipelines or services processing untrusted model assets face elevated exposure.
Меры по смягчению последствий
To mitigate this issue, avoid processing .traineddata files from untrusted or unverified sources with Tesseract. Restricting the application's input to only trusted .traineddata files can prevent exploitation of this vulnerability.
Дополнительная информация
Статус:
EPSS
7.8 High
CVSS3
Связанные уязвимости
Tesseract is an open source OCR engine. In version 5.5.3 and earlier, prior .traineddata hardening added bounds checks to NetworkIO::CopyTimeStepGeneral and NetworkIO::Randomize in src/lstm/networkio.cpp but left NetworkIO::WriteTimeStepPart and NetworkIO::AddTimeStepPart unchecked. In LSTM::Forward in src/lstm/lstm.cpp, source_ is sized from the independently deserialized na_ field while the WriteTimeStepPart count is ns_, which comes from the CI gate WeightMatrix dim1() value. A crafted NT_LSTM layer can make ns_ much larger than na_, causing a heap out-of-bounds write during the first recognition step on the default LSTM engine and resulting in heap corruption, a crash, or potentially controlled corruption. No fixed release is available as of this review.
Tesseract is an open source OCR engine. In version 5.5.3 and earlier, prior .traineddata hardening added bounds checks to NetworkIO::CopyTimeStepGeneral and NetworkIO::Randomize in src/lstm/networkio.cpp but left NetworkIO::WriteTimeStepPart and NetworkIO::AddTimeStepPart unchecked. In LSTM::Forward in src/lstm/lstm.cpp, source_ is sized from the independently deserialized na_ field while the WriteTimeStepPart count is ns_, which comes from the CI gate WeightMatrix dim1() value. A crafted NT_LSTM layer can make ns_ much larger than na_, causing a heap out-of-bounds write during the first recognition step on the default LSTM engine and resulting in heap corruption, a crash, or potentially controlled corruption. No fixed release is available as of this review.
Tesseract is an open source OCR engine. In version 5.5.3 and earlier, ...
Уязвимость функций NetworkIO::CopyTimeStepGeneral и NetworkIO::Randomize системы оптического распознавания символов Tesseract, позволяющая нарушителю оказать воздействие на конфиденциальность, целостность и доступность защищаемой информации
EPSS
7.8 High
CVSS3