def error_quantization(image, error, maxvals, m=None):
"""
(Task 4): Given image as a list of values post-predictive coding, and m,
perform uniform OR non-uniform quantization of the error into m bins.
"""
if m is None:
e1 = int_seq_to_bytearray(error[0], maxvals[0])
e2 = int_seq_to_bytearray(error[1], maxvals[1])
e3 = int_seq_to_bytearray(error[2], maxvals[2])
return e1, e2, e3
e1, e2, e3 = error
e1New = error_quantize(e1, m, maxvals[0])
e2New = error_quantize(e2, m, maxvals[1])
e3New = error_quantize(e3, m, maxvals[2])
return e1New, e2New, e3New
def error_quantization(image, error, maxvals, m=None):
"""
(Task 4): Given image as a list of values post-predictive coding, and m,
perform uniform OR non-uniform quantization of the error into m bins.
"""
if m is None:
e1 = int_seq_to_bytearray(error[0], maxvals[0])
e2 = int_seq_to_bytearray(error[1], maxvals[1])
e3 = int_seq_to_bytearray(error[2], maxvals[2])
return e1, e2, e3
e1, e2, e3 = error
e1New = error_quantize(e1, m, maxvals[0])
e2New = error_quantize(e2, m, maxvals[1])
e3New = error_quantize(e3, m, maxvals[2])
return e1New, e2New, e3New
def error_quantize(channel, num_bins, maxval):
if not num_bins:
return channel
# Get amount covered by each bin
bin_size = int(math.ceil(maxval * 2 / num_bins))
# Get bin for each value in channel
bin_vals = [int((i + maxval) / bin_size) for i in channel]
return int_seq_to_bytearray(bin_vals, num_bins - 1)
def error_quantize(channel, num_bins, maxval):
if not num_bins:
return channel
# Get amount covered by each bin
bin_size = int(math.ceil(maxval * 2 / num_bins))
# Get bin for each value in channel
bin_vals = [int((i + maxval) / bin_size) for i in channel]
return int_seq_to_bytearray(bin_vals, num_bins-1)
