For image provision module is used for image classification, segmentation, and localization tasks. We support any image format that can be decoded with OpenCV.

The complete table of configuration parameters is shown below:

Parameter Default Description
height (uint) Required Height of provisioned image (pixels)
width (uint) Required Width of provisioned image (pixels)
name (string) “” Name prepended to the output buffer name
output_type (string) “uint8_t” Output data type.
channels (uint) 3 Number of channels in input image
channel_major (bool) True Load the pixel buffer in channel major order (that is, all pixels from blue channel contiguous, followed by all pixels from green channel, followed by all pixels from the red channel). The alternative is to have the color channels for each pixel located adjacent to each other (b1g1r1b2g2r2 rather than b1b2g1g2r1r2).
seed (int) 0 Random seed

The buffers provisioned to the model are:

Buffer Name Shape Description
image (N, C, H, W) or (N, H, W, C) Extracted and transformed image, where C = channels, H = height, W = width, and N = bsz (the batch size). Layout is dependent on whether channel_major mode is true or not