diff --git a/skimage/transform/rescale.py b/skimage/transform/rescale.py index 4ec4ab99..b7164994 100644 --- a/skimage/transform/rescale.py +++ b/skimage/transform/rescale.py @@ -10,21 +10,17 @@ def downsample(image, factors, method='sum'): f1 = factors[1] if (f0 - int(f0) != 0) or (f1 - int(f1) != 0): - print "Use resample for non-integer downsampling" + print "Use resample() for non-integer downsampling" return - cropped = image[: is0 - (is0 % f0), : is1 - (is1 % f1)] + cropped = image[:is0 - (is0 % f0), :is1 - (is1 % f1)] out = np.zeros((cropped.shape[0] / f0, cropped.shape[1] / f1)) + for i in range(cropped.shape[0]): + for j in range(cropped.shape[1]): + out[int(i / f0)][int(j / f1)] += cropped[i][j] if method == 'sum': - for i in range(cropped.shape[0]): - for j in range(cropped.shape[1]): - out[int(i / f0)][int(j / f1)] += cropped[i][j] return out - - if method == 'mean': - for i in range(cropped.shape[0]): - for j in range(cropped.shape[1]): - out[int(i / f0)][int(j / f1)] += cropped[i][j] + else: return out / float(f0 * f1) @@ -36,18 +32,15 @@ def upsample(image, factors, method='divide'): f1 = factors[1] if (f0 - int(f0) != 0) or (f1 - int(f1) != 0): - print "Use resample for non-integer upsampling" + print "Use resample() for non-integer upsampling" return out = np.zeros((f0 * image.shape[0], f1 * image.shape[1])) - if method == 'divide': - for i in range(out.shape[0]): - for j in range(out.shape[1]): - out[i][j] = (image[i / f0][j / f1]) - return out / float(f0 * f1) - if method == 'uniform': - for i in range(out.shape[0]): - for j in range(out.shape[1]): - out[i][j] = (image[i / f0][j / f1]) + for i in range(out.shape[0]): + for j in range(out.shape[1]): + out[i][j] = (image[i / f0][j / f1]) + if method == 'divide': + return out / float(f0 * f1) + else: return out