MdlocrFinetune

Board	Supported
Host System	Yes
V4L2	Yes
Clarity UHD	Yes
Concord PoE	No
GenTL	Yes
GevIQ	Yes
GigE Vision	Yes
Indio	No
Iris GTX	Yes
Radient eV-CL	Yes
Rapixo CL	Yes
Rapixo CoF	Yes
Rapixo CXP	Yes
USB3 Vision	Yes

Fine-tune a Deep Learning OCR read context on a dataset.

Syntax

void MdlocrFinetune(
    AIL_ID    FinetuneContextDlocrId,  //in
    AIL_ID    ReadContextDlocrId,      //in
    AIL_ID    DlocrDatasetId,          //in
    AIL_ID    FinetuneResultDlocrId,   //out
    AIL_INT64 ControlFlag              //in
)

Description

This function uses a Deep Learning OCR dataset to fine-tune a Deep Learning OCR read context. Fine-tuning results are stored in a Deep Learning OCR result buffer. To get the results, use MdlocrGetResult.

The fine-tuning process is time-consuming and heavily relies on the accuracy of the dataset's annotations. That is, the extent to which fine-tuning improves the read operation depends on the image quality of the characters or strings in the source dataset. Training time can vary considerably depending on the complexity of the images, the available hardware, and the accuracy required. Images can be added to the dataset, using MdlocrAddImageToDataset. You can remove the last image that was added to the dataset, using MdlocrDeleteLastImageFromDataset. You can export a dataset to the disk, using MdlocrExport or restore a dataset from the disk, using MdlocrImport.

Before performing a fine-tuning operation, the Deep Learning OCR read and fine-tuning contexts must have been previously allocated, usingMdlocrAlloc and be in a preprocessed state (MdlocrPreprocess).

You can manually stop the fine-tuning operation, using MdlocrControl with.

Parameters

FinetuneContextDlocrId (in, AIL_ID)

Specifies the identifier of the Deep Learning OCR fine-tuning context to use for the fine-tuning operation. The context must have been previously allocated on the required system using MdlocrAlloc, and preprocessed using MdlocrPreprocess.

ReadContextDlocrId (in, AIL_ID)

Specifies the identifier of the Deep Learning OCR read context to use for the fine-tuning operation. The context must have been previously allocated on the required system using MdlocrAlloc, and preprocessed using MdlocrPreprocess.

DlocrDatasetId (in, AIL_ID)

Specifies the identifier of the Deep Learning OCR dataset to use for the fine-tuning operation. The dataset must have been previously allocated on the required system using MdlocrAllocDataset.

FinetuneResultDlocrId (out, AIL_ID)

Specifies the identifier of the Deep Learning OCR fine-tuning result buffer in which to write the results of the fine-tuning operation. The fine-tuning result buffer must have been previously allocated on the required system using MdlocrAllocResult.

ControlFlag (in, AIL_INT64)

Reserved for future expansion and must be set to M_DEFAULT.