Understanding Size Discrepancy between Raw Array and Image from Raspberry Pi Camera

Question

I am using a Raspberry Pi Compute Module 4 with an Arducam IMX230 camera to capture raw images. The camera is configured for a specific sensor mode, but the resulting numpy array from the raw capture does not match the expected resolution. Here is the sensor mode configuration:

config = {
'format': 'SRGGB10_CSI2P',
'unpacked': 'SRGGB10',
'bit_depth': 10,
'size': (5344, 4012),
'fps': 9.0,
'crop_limits': (0, 0, 5344, 4016),
'exposure_limits': (536, 999871, None)
}

The expected resolution based on the 'size' parameter is (5344, 4012), but the numpy array I get after capturing an image is of shape (6688, 4012).

This only happens when I capture arrays, when I capture images I get the correct shapes.

Can anyone help explain why there is a discrepancy in the image resolution?