| |
| |
| nmb_of_modules=40; |
| nmb_of_module_subsets=2; |
| channels_names={'R','G','B','RGg1','RBg1','GBg1','RGg2','RBg2','GBg2','RB','RG','GB','eRGB','BW','X','Y','Z'}; |
| feature_RGB=[1 0 0 |
| 0 1 0 |
| 0 0 1 |
| 0.618 0.382 0 |
| 0.618 0 0.382 |
| 0 0.618 0.382 |
| 0.382 0.618 0 |
| 0.382 0 0.618 |
| 0 0.382 0.618 |
| 0.5 0.5 0 |
| 0.5 0 0.5 |
| 0 0.5 0.5 |
| 1/3 1/3 1/3 |
| 0.299 0.587 0.114 |
| 0.4125 0.3576 0.1804 |
| 0.2126 0.7152 0.0722 |
| 0.0193 0.1192 0.9502]; |
| nmb_of_colors=length(channels_names); |
| patch=0; |
| nmb_of_labs_per_module=25; |
| cross_entropy=1; |
|
|
| param.image_size=[64,64]; |
| param.downsizing=2; |
| param.x_trim=1; |
| param.y_trim=1; |
| param.compute_decimal_place=4; |
| param.dwnsz_on=1; |
|
|
| param.patch=patch; |
| param.nmb_of_modules=nmb_of_modules; |
| param.nmb_of_module_subsets=nmb_of_module_subsets; |
|
|
| param.cross_entropy=cross_entropy; |
| param.nmb_of_labs_per_module=nmb_of_labs_per_module; |
|
|
| param.channels_names=channels_names; |
| param.nmb_of_colors=nmb_of_colors; |
| param.feature_RGB=feature_RGB; |
|
|
| parfor module=1:nmb_of_modules |
| data_param=[]; |
| for subset=1:nmb_of_module_subsets |
| reportname1 = sprintf('Transformed_IN1k_Data/Modularized_Data_for_SGD/modularized_data_patch_%d_module_%d_subset_%d_for_%d_labels_per_module.mat', ... |
| patch,module,subset,nmb_of_labs_per_module); |
| |
| data_load=load(reportname1); |
| data_0=data_load.data; |
| output=fun_transform_data_rgbfeatures(data_0,param); |
| data=output.transformed_image; |
| data_param.mnsv=output.mnsv; |
| data_param.maxsv=output.maxsv; |
| data_param.ipvsz=output.ipvsz; |
| labels=data_load.labels; |
| label_ids=data_load.label_ids; |
| label_table=data_load.label_table; |
| out=fun_save_transformed_data(patch, module, subset,nmb_of_labs_per_module,data,labels,label_ids,label_table,data_param); |
| end |
| module |
| end |
| |
