BarcodeRecognizer::InitializationParams

The initialization parameters for the barcode recognizer. More...

#include <ScanbotSDK.hpp>

Public Attributes
scanbotsdk_barcode_engine_mode_t	engineMode = SCANBOTSDK_BARCODE_ENGINE_MODE_BALANCED
	The barcode recognition engine mode.
std::vector< scanbotsdk_barcode_format_t >	formats
	The barcode formats to detect. If empty, all formats will be recognized.
bool	useTensorRT = false
	If true, the barcode recognizer will use the TensorRT backend for GPU acceleration.
int	tensorrtMaxInputWidth = 0
	The maximum input width in pixels for TensorRT. If 0, the default value will be used.
int	tensorrtMaxInputHeight = 0
	The maximum input height in pixels for TensorRT. If 0, the default value will be used.
size_t	tensorrtMaxWorkspaceSize = 0
	The maximum workspace size in bytes for TensorRT. If 0, the default value will be used.

Detailed Description

The initialization parameters for the barcode recognizer.

Member Data Documentation

useTensorRT

bool useTensorRT = false

If true, the barcode recognizer will use the TensorRT backend for GPU acceleration.

When enabling TensorRT, use the smallest input image size that fits your use case. The bigger the image size, the more memory is required at runtime. It takes around 3 minutes to build the TensorRT engine for the first time. The engine is then cached on disk and reused. The cached engine can be copied to other devices with the same device model, Jetpack version, CUDA version and TensorRT version.

---

The documentation for this struct was generated from the following file:

• c/include/ScanbotSDK.hpp