Open SaashaJoshi opened 7 months ago
Do INEQR
and GQIR
require different implementations?
INEQR
represents images of 2^n1 x 2^n2
dimensions, whereas, GQIR
represents images of arbitrary dimensions H x W
. According to this, INEQR
could very much be a subset of GQIR
.
Answer: Yes, with minor modifications, there is a need for different implementations for INEQR
and GQIR
.
GQIR
needs to modify the pixel_vals
matrix to include dummy or None
values for the redundancy pixels. For example, if the image is of size 1x3
, a total of 2 + 4 = 3
qubits will be required to represent the image on the quantum circuit. However, the pixel_vals
matrix should be transformed from something like this [[1, 2, 3]]
to something like this [[1, 2, 3, None] [None, None, None, None]]
.
Image taken from:
N. Jiang, J. Wang, and Y. Mu, “Quantum image scaling up based on nearest-neighbor interpolation with integer scaling ratio,” Quantum Information Processing, vol. 14, no. 11, pp. 4001–4026, Aug. 2015, doi: https://doi.org/10.1007/s11128-015-1099-5
The pipeline currently has 2 image representation algorithms:
Important QIR methods that need to be added: