Vision Apps User Guide

Data Fields

vx_float32 mean_pixel [3]
 
vx_float32 scale_val [3]
 
vx_int32 pad_pixel [4]
 
vx_int32 tidl_8bit_16bit_flag
 
vx_int32 ip_rgb_or_yuv
 
vx_int32 color_conv_flag
 
vx_uint32 clear_count
 
vx_int32 skip_flag
 

Field Documentation

◆ mean_pixel

vx_float32 tivxImgPreProcParams::mean_pixel[3]

Mean pixel for each output plane which needs to be subtracted

◆ scale_val

vx_float32 tivxImgPreProcParams::scale_val[3]

Scale value for each output plane which needs to be multiplied

◆ pad_pixel

vx_int32 tivxImgPreProcParams::pad_pixel[4]

Pad value for output planes for each side of image

◆ tidl_8bit_16bit_flag

vx_int32 tivxImgPreProcParams::tidl_8bit_16bit_flag

To enable 8 bit or 16 bit input/output tensor for tidl. 0 means 8 bit, 1 means 16 bit

◆ ip_rgb_or_yuv

vx_int32 tivxImgPreProcParams::ip_rgb_or_yuv

Falg to indicate input is RGB or YUV. 0 means RGB otherwise YUV

◆ color_conv_flag

vx_int32 tivxImgPreProcParams::color_conv_flag

color conversion flag for seperating the plane as peer the model requirement

◆ clear_count

vx_uint32 tivxImgPreProcParams::clear_count

Number of times to clear the output buffer

◆ skip_flag

vx_int32 tivxImgPreProcParams::skip_flag

Skip processing