==============================================================
EVAL_NMI_CPS_WARP: get similarity score between CPS warped image and
reference. (Evaluate the clamped-plate spline warp using
mean-squared-differences)
Code written by Katherine Smith, 2003
GENERAL
[score, warped_points] = eval_nmi_cps_warp(params, warp_centre,
r, ref_points, start_image, ref_image)
INPUT/S
-params:
The paramters of the warps.
params(1): warp centre
params(2): r
params(3): d
-warp centre:
Not in parameters already??
-r:
Not in parameters already??
-start_points_vec:
Points before CPS warp.
-start_image_vec:
Image before CPS warp.
-ref_image_vec:
The reference image to be compared to.
OUTPUT/S
-score:
The similarity score w.r.t. the reference image.
(MSD)
-warped_points:
The position of points after warping.
PENDING WORK
-
KNOWN BUG/S
-None.
COMMENT/S
-
RELATED FUNCTION/S
EVAL_MI_CPS_WARP
ABOUT
-Created: November 23rd, 2003
-Last update: Novermber 27th, 2003
-Revision: 0.0.2
-Author: R. S. Schestowitz, University of Manchester
==============================================================0001 function [score, warped_points] = eval_nmi_cps_warp(params, warp_centre, r, ref_points, start_image, ref_image, number_of_bins) 0002 % ============================================================== 0003 % EVAL_NMI_CPS_WARP: get similarity score between CPS warped image and 0004 % reference. (Evaluate the clamped-plate spline warp using 0005 % mean-squared-differences) 0006 % 0007 % Code written by Katherine Smith, 2003 0008 % 0009 % GENERAL 0010 % 0011 % [score, warped_points] = eval_nmi_cps_warp(params, warp_centre, 0012 % r, ref_points, start_image, ref_image) 0013 % 0014 % INPUT/S 0015 % 0016 % -params: 0017 % The paramters of the warps. 0018 % 0019 % params(1): warp centre 0020 % params(2): r 0021 % params(3): d 0022 % 0023 % -warp centre: 0024 % Not in parameters already?? 0025 % 0026 % -r: 0027 % Not in parameters already?? 0028 % 0029 % -start_points_vec: 0030 % Points before CPS warp. 0031 % 0032 % -start_image_vec: 0033 % Image before CPS warp. 0034 % 0035 % -ref_image_vec: 0036 % The reference image to be compared to. 0037 % 0038 % OUTPUT/S 0039 % 0040 % -score: 0041 % The similarity score w.r.t. the reference image. 0042 % (MSD) 0043 % 0044 % -warped_points: 0045 % The position of points after warping. 0046 % 0047 % 0048 % PENDING WORK 0049 % 0050 % - 0051 % 0052 % KNOWN BUG/S 0053 % 0054 % -None. 0055 % 0056 % COMMENT/S 0057 % 0058 % - 0059 % 0060 % RELATED FUNCTION/S 0061 % 0062 % EVAL_MI_CPS_WARP 0063 % 0064 % ABOUT 0065 % 0066 % -Created: November 23rd, 2003 0067 % -Last update: Novermber 27th, 2003 0068 % -Revision: 0.0.2 0069 % -Author: R. S. Schestowitz, University of Manchester 0070 % ============================================================== 0071 0072 d = params; 0073 % copy parameters 0074 start_points = ref_points; 0075 max_d = 0.67; 0076 % maximum parameter value allowed?? 0077 0078 if (any((abs(d) > max_d)) | any(((warp_centre-r <-1) | (warp_centre+r > 1)))) 0079 0080 % ensure for all paramters p,-0.67<p<0.67 and also 0081 % warp center and radius are not out of normalisation 0082 % range 1. All must be in bounds 0083 0084 score = realmax; 0085 % if data is invalid, set to maximum <Real> value 0086 0087 else 0088 % if inputs okay 0089 for i=1:size(params,2) 0090 % for all parameters 0091 warped_points = cps_warp_1d(start_points,warp_centre(i),r(i),d(i)); 0092 % apply warp 0093 start_points = warped_points; 0094 % proceed with warped point to next iteration 0095 end 0096 warped_image = interp1(ref_points,start_image,warped_points, 'linear',0); 0097 % record warped image 0098 score = nmi(ref_image, warped_image, number_of_bins); 0099 % score according to msd 0100 end