Skip to main content
Version: 3.20.2

Processing Block

Requirements​

Create Processing Blocks​

The ProcessingBlock object is used to detect faces and perform image quality assessment. Block Type Modification

Procedure for using the object ProcessingBlock:

  1. Create an object ProcessingBlock
ProcessingBlock processingBlock = service.createProcessingBlock({
      "unit_type": "<Block Type>",
      "modification": "<Modification>",
      "@sdk_path": "<path to Face SDK>"
    });
  1. Get an image and convert it to RGB format
  2. Create a Context object with the following data
final Context data = service.createContext({
      "objects": [],
      "image": {
        "blob": <byte array with image in RGB format>,
        "dtype": "uint8_t",
        "format": "NDARRAY",
        "shape": [<image height in pixels>, <image width in pixels>, 3]
      }
});
  1. Call the process method of ProcessingBlock
Context result = processingBlock.process(data);
  1. Output data is in result["objects"]
    Example of data usage

    Getting the coordinates of the top left and bottom right points of the bbox

for (int i = 0; i < objects.len(); i++) {
      Context object = objects[i];
      Context bbox = object["bbox"];

      // Top left point of bbox
      double x1 = bbox[0].get_value() * imageWidth;
      double y1 = bbox[1].get_value() * imageHeight;
      
      // Bottom right point of bbox
      double x2 = bbox[2].get_value() * imageWidth;
      double y2 = bbox[3].get_value() * imageHeight;
}

1. Creating ProcessingBlock FaceDetector object​

FaceDetector is used to detect faces in an image

ProcessingBlock faceDetector = service.createProcessingBlock({
      "unit_type": "FACE_DETECTOR",
      "modification": "uld",
      "@sdk_path": "<path to Face SDK>",
      "min_size": 50
});

2. Creating a ProcessingBlock FaceFitter object​

FaceFitter is used to find key points of faces in an image
Some ProcessingBlocks require the FaceDetector + FaceFitter bundle to work

ProcessingBlock fitter = service.createProcessingBlock({
      "unit_type": "FACE_FITTER",
      "modification": "tddfa_faster",
      "@sdk_path": "<path to Face SDK>"
});

3. Creating ProcessingBlock QualityAssessment object​

QualityAssessment is used to assess the quality of an image using multiple parameters
It requires FaceFitter to work

ProcessingBlock qualityAssessment = service.createProcessingBlock({
      "unit_type": "QUALITY_ASSESSMENT_ESTIMATOR",
      "modification": "assessment",
      "config_name": "quality_assessment.xml",
      "@sdk_path": "<path to Face SDK>"
});

4. Creating ProcessingBlock QualityEstimator object​

The QualityEstimator is used to evaluate image quality

ProcessingBlock qualityEstimator = service.createProcessingBlock({
      "unit_type": "QUALITY_ASSESSMENT_ESTIMATOR", 
      "modification": "estimation", 
      "@sdk_path": "<path to Face SDK>"
});