vpi.ConfidenceType

class vpi.ConfidenceType

Defines the way the confidence values are computed.

This computation needs disparity difference from left to right (D_lr) and from right to left (D_rl), and potentially the maximum disparity (MAX_DISP). Lower confidence values (than the given confidence threshold parameter) make the output disparity invalid.

Enumeration values

vpi.ConfidenceType.ABSOLUTE	The U16 confidence value of a pixel is given by: CUDA backend: [ 1 - abs(D_lr - D_rl) / MAX_DISP ] * 0xFFFF; Other backends: [ 1 - abs(D_lr - D_rl) / 16 ] * 0xFF00
vpi.ConfidenceType.RELATIVE	The U16 confidence value of a pixel is given by: CUDA backend: [ 1 - abs(D_lr - D_rl) / D_lr ] * 0xFFFF; Other backends: [ 1 - abs(D_lr - D_rl) / 16 ] * 0xFF00
vpi.ConfidenceType.INFERENCE	The U16 confidence value of a pixel is a number between 0:UINT16_MAX representing 0:100% confidence in the disparity value (only available for OFA+PVA+VIC backend)