open source pkg v1

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Prepare_HOG_AU_data_generic.m

@@ -0,0 +1,106 @@
+function [data_train, labels_train, data_devel, labels_devel, raw_devel, PC, means_norm, stds_norm, vid_ids_devel_string, valid_ids_devel] = ...
+    Prepare_HOG_AU_data_generic(train_users, devel_users, au_train, rest_aus, fera_dir, features_dir)
+
+%%
+addpath(genpath('../../data extraction/'));
+
+% First extracting the labels
+[ labels_train, valid_ids_train, filenames ] = extract_FERA2011_labels(fera_dir, train_users, au_train);
+
+[ labels_other, ~, ~ ] = extract_FERA2011_labels(fera_dir, train_users, rest_aus);
+labels_other = cat(1, labels_other{:});
+
+% Reading in the HOG data (of only relevant frames)
+[train_appearance_data, valid_ids_train_hog, vid_ids_train_string] = Read_HOG_files(train_users, features_dir);
+
+[train_geom_data] = Read_geom_files(train_users, features_dir);
+
+% Subsample the data to make training quicker
+labels_train = cat(1, labels_train{:});
+valid_ids_train = logical(cat(1, valid_ids_train{:}));
+
+reduced_inds = false(size(labels_train,1),1);
+reduced_inds(labels_train == 1) = true;
+
+% make sure the same number of positive and negative samples is taken
+pos_count = sum(labels_train == 1);
+neg_count = sum(labels_train == 0);
+
+num_other = floor(pos_count / (size(labels_other, 2)));
+
+inds_all = 1:size(labels_train,1);
+
+for i=1:size(labels_other, 2)+1
+   
+    if(i > size(labels_other, 2))
+        % fill the rest with a proportion of neutral
+        inds_other = inds_all(sum(labels_other,2)==0 & ~labels_train );   
+        num_other_i = min(numel(inds_other), pos_count - sum(labels_train(reduced_inds,:)==0));     
+    else
+        % take a proportion of each other AU
+        inds_other = inds_all(labels_other(:, i) & ~labels_train );      
+        num_other_i = min(numel(inds_other), num_other);        
+    end
+    inds_other_to_keep = inds_other(round(linspace(1, numel(inds_other), num_other_i)));
+    reduced_inds(inds_other_to_keep) = true;
+    
+end
+
+% Remove invalid ids based on CLM failing or AU not being labelled
+reduced_inds(~valid_ids_train) = false;
+reduced_inds(~valid_ids_train_hog) = false;
+
+labels_other = labels_other(reduced_inds, :);
+labels_train = labels_train(reduced_inds,:);
+train_appearance_data = train_appearance_data(reduced_inds,:);
+train_geom_data = train_geom_data(reduced_inds,:);
+vid_ids_train_string = vid_ids_train_string(reduced_inds,:);
+
+%% Extract devel data
+
+% First extracting the labels
+[ labels_devel, valid_ids_devel, filenames_devel ] = extract_FERA2011_labels(fera_dir, devel_users, au_train);
+
+% Reading in the HOG data (of only relevant frames)
+[devel_appearance_data, valid_ids_devel_hog, vid_ids_devel_string] = Read_HOG_files(devel_users, features_dir);
+
+[devel_geom_data] = Read_geom_files(devel_users, features_dir);
+
+labels_devel = cat(1, labels_devel{:});
+
+% Peforming zone specific masking
+if(au_train < 8 || au_train == 43 || au_train == 45) % upper face AUs ignore bottom face
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_upper.mat';
+    load(pca_file);
+elseif(au_train > 9) % lower face AUs ignore upper face and the sides
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_lower.mat';
+    load(pca_file);
+elseif(au_train == 9) % Central face model
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_rigid.mat';
+    load(pca_file);
+end
+
+% Grab all data for validation as want good params for all the data
+raw_devel = cat(2, devel_appearance_data, devel_geom_data);
+
+devel_appearance_data = bsxfun(@times, bsxfun(@plus, devel_appearance_data, -means_norm), 1./stds_norm);
+train_appearance_data = bsxfun(@times, bsxfun(@plus, train_appearance_data, -means_norm), 1./stds_norm);
+
+data_train = train_appearance_data * PC;
+data_devel = devel_appearance_data * PC;
+
+data_train = cat(2, data_train, train_geom_data);
+data_devel = cat(2, data_devel, devel_geom_data);
+
+PC_n = zeros(size(PC)+size(train_geom_data, 2));
+PC_n(1:size(PC,1), 1:size(PC,2)) = PC;
+PC_n(size(PC,1)+1:end, size(PC,2)+1:end) = eye(size(train_geom_data, 2));
+PC = PC_n;
+
+means_norm = cat(2, means_norm, zeros(1, size(train_geom_data,2)));
+stds_norm = cat(2, stds_norm, ones(1, size(train_geom_data,2)));
+
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Prepare_HOG_AU_data_generic_dynamic.m

@@ -0,0 +1,106 @@
+function [data_train, labels_train, data_devel, labels_devel, raw_devel, PC, means_norm, stds_norm, vid_ids_devel_string] = ...
+    Prepare_HOG_AU_data_generic_dynamic(train_users, devel_users, au_train, rest_aus, semaine_dir, features_dir)
+
+%%
+addpath(genpath('../../data extraction/'));
+
+% First extracting the labels
+[ labels_train, filenames, vid_ids_train ] = extract_FERA2011_labels(semaine_dir, train_users, au_train);
+
+[ labels_other, ~, ~ ] = extract_FERA2011_labels(semaine_dir, train_users, rest_aus);
+labels_other = cat(1, labels_other{:});
+
+% Reading in the HOG data (of only relevant frames)
+[train_appearance_data, valid_ids_train_hog, vid_ids_train_string] = Read_HOG_files_dynamic(train_users, features_dir);
+
+[train_geom_data] = Read_geom_files_dynamic(train_users, features_dir);
+
+% Subsample the data to make training quicker
+labels_train = cat(1, labels_train{:});
+filenames = logical(cat(1, filenames{:}));
+
+reduced_inds = false(size(labels_train,1),1);
+reduced_inds(labels_train == 1) = true;
+
+% make sure the same number of positive and negative samples is taken
+pos_count = sum(labels_train == 1);
+neg_count = sum(labels_train == 0);
+
+num_other = floor(pos_count / (size(labels_other, 2)));
+
+inds_all = 1:size(labels_train,1);
+
+for i=1:size(labels_other, 2)+1
+   
+    if(i > size(labels_other, 2))
+        % fill the rest with a proportion of neutral
+        inds_other = inds_all(sum(labels_other,2)==0 & ~labels_train );   
+        num_other_i = min(numel(inds_other), pos_count - sum(labels_train(reduced_inds,:)==0));     
+    else
+        % take a proportion of each other AU
+        inds_other = inds_all(labels_other(:, i) & ~labels_train );      
+        num_other_i = min(numel(inds_other), num_other);        
+    end
+    inds_other_to_keep = inds_other(round(linspace(1, numel(inds_other), num_other_i)));
+    reduced_inds(inds_other_to_keep) = true;
+    
+end
+
+% Remove invalid ids based on CLM failing or AU not being labelled
+reduced_inds(~filenames) = false;
+reduced_inds(~valid_ids_train_hog) = false;
+
+labels_other = labels_other(reduced_inds, :);
+labels_train = labels_train(reduced_inds,:);
+train_appearance_data = train_appearance_data(reduced_inds,:);
+train_geom_data = train_geom_data(reduced_inds,:);
+vid_ids_train_string = vid_ids_train_string(reduced_inds,:);
+
+%% Extract devel data
+
+% First extracting the labels
+[ labels_devel, valid_ids_devel, vid_ids_devel ] = extract_FERA2011_labels(semaine_dir, devel_users, au_train);
+
+% Reading in the HOG data (of only relevant frames)
+[devel_appearance_data, valid_ids_devel_hog, vid_ids_devel_string] = Read_HOG_files_dynamic(devel_users, features_dir);
+
+[devel_geom_data] = Read_geom_files_dynamic(devel_users, features_dir);
+
+labels_devel = cat(1, labels_devel{:});
+
+% Peforming zone specific masking
+if(au_train < 8 || au_train == 43 || au_train == 45) % upper face AUs ignore bottom face
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_upper.mat';
+    load(pca_file);
+elseif(au_train > 9) % lower face AUs ignore upper face and the sides
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_lower.mat';
+    load(pca_file);
+elseif(au_train == 9) % Central face model
+    % normalise the data
+    pca_file = '../../pca_generation/generic_face_rigid.mat';
+    load(pca_file);
+end
+     
+% Grab all data for validation as want good params for all the data
+raw_devel = cat(2, devel_appearance_data, devel_geom_data);
+
+devel_appearance_data = bsxfun(@times, bsxfun(@plus, devel_appearance_data, -means_norm), 1./stds_norm);
+train_appearance_data = bsxfun(@times, bsxfun(@plus, train_appearance_data, -means_norm), 1./stds_norm);
+
+data_train = train_appearance_data * PC;
+data_devel = devel_appearance_data * PC;
+
+data_train = cat(2, data_train, train_geom_data);
+data_devel = cat(2, data_devel, devel_geom_data);
+
+PC_n = zeros(size(PC)+size(train_geom_data, 2));
+PC_n(1:size(PC,1), 1:size(PC,2)) = PC;
+PC_n(size(PC,1)+1:end, size(PC,2)+1:end) = eye(size(train_geom_data, 2));
+PC = PC_n;
+
+means_norm = cat(2, means_norm, zeros(1, size(train_geom_data,2)));
+stds_norm = cat(2, stds_norm, ones(1, size(train_geom_data,2)));
+
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Read_HOG_files.m

@@ -0,0 +1,86 @@
+function [hog_data, valid_data, vid_id] = Read_HOG_files(users, hog_data_dir)
+
+    hog_data = [];
+    vid_id = {};
+    valid_data = [];
+    
+    feats_filled = 0;
+
+    for i=1:numel(users)
+        
+        hog_file = [hog_data_dir, '/au_training_' users{i} '.hog'];
+        
+        f = fopen(hog_file, 'r');
+                          
+        curr_data = [];
+        curr_ind = 0;
+        
+        while(~feof(f))
+                        
+            if(curr_ind == 0)
+                num_cols = fread(f, 1, 'int32');
+                if(isempty(num_cols))
+                    break;
+                end
+
+                num_rows = fread(f, 1, 'int32');
+                num_chan = fread(f, 1, 'int32');
+
+                curr_ind = curr_ind + 1;            
+
+                % preallocate some space
+                if(curr_ind == 1)
+                    curr_data = zeros(5000, 1 + num_rows * num_cols * num_chan);
+                    num_feats =  1 + num_rows * num_cols * num_chan;
+                end
+
+                if(curr_ind > size(curr_data,1))
+                    curr_data = cat(1, curr_data, zeros(6000, 1 + num_rows * num_cols * num_chan));
+                end
+                feature_vec = fread(f, [1, 1 + num_rows * num_cols * num_chan], 'float32');
+                curr_data(curr_ind, :) = feature_vec;
+            else
+            
+                % Reading in batches of 5000
+                
+                feature_vec = fread(f, [4 + num_rows * num_cols * num_chan, 5000], 'float32');
+                feature_vec = feature_vec(4:end,:)';
+                
+                num_rows_read = size(feature_vec,1);
+                
+                curr_data(curr_ind+1:curr_ind+num_rows_read,:) = feature_vec;
+                
+                curr_ind = curr_ind + size(feature_vec,1);
+                
+            end
+                        
+        end
+        
+        fclose(f);
+        
+        curr_data = curr_data(1:curr_ind,:);
+        vid_id_curr = cell(curr_ind,1);
+        vid_id_curr(:) = users(i);
+        
+        vid_id = cat(1, vid_id, vid_id_curr);
+        
+        % Assume same number of frames per video
+        if(i==1)
+            hog_data = zeros(curr_ind*numel(users), num_feats);
+        end
+        
+        if(size(hog_data,1) < feats_filled+curr_ind)
+           hog_data = cat(1, hog_data, zeros(size(hog_data,1), num_feats));
+        end
+        
+        hog_data(feats_filled+1:feats_filled+curr_ind,:) = curr_data;
+        
+        feats_filled = feats_filled + curr_ind;
+        
+    end
+    if(numel(users) > 0)        
+        valid_data = hog_data(1:feats_filled,1) > 0;
+        hog_data = hog_data(1:feats_filled,2:end);
+    end
+    
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Read_HOG_files_dynamic.m

@@ -0,0 +1,133 @@
+function [hog_data, valid_data, vid_id] = Read_HOG_files_dynamic(users, hog_data_dir)
+
+    hog_data = [];
+    vid_id = {};
+    valid_data = [];
+    
+    feats_filled = 0;
+    
+    user1 = {'train_001', 'train_002', 'train_003', 'train_004', 'train_005',...
+            'train_006', 'train_007', 'train_008', 'train_009', 'train_010',...
+            'train_011', 'train_012', 'train_013', 'train_014', 'train_015',...
+            'train_016', 'train_017', 'train_018'};
+    
+    user2 = {'train_019', 'train_020', 'train_021', 'train_022', 'train_023',...
+             'train_024', 'train_025', 'train_026', 'train_027', 'train_028',...
+             'train_029', 'train_030', 'train_031', 'train_032'};
+        
+    user3 = {'train_033', 'train_034', 'train_035', 'train_036', 'train_037',...
+             'train_038', 'train_039', 'train_040', 'train_041'};
+         
+    user4 = {'train_042', 'train_043', 'train_044', 'train_045', 'train_046',...
+             'train_047', 'train_048', 'train_049', 'train_050', 'train_051',...
+             'train_052', 'train_053', 'train_054', 'train_055', 'train_056'};
+    
+    user5 = {'train_057', 'train_058', 'train_059', 'train_060', 'train_061',...
+             'train_062', 'train_063'};
+         
+    user6 = {'train_064', 'train_065', 'train_066', 'train_067', 'train_068',...
+             'train_069', 'train_070', 'train_071', 'train_072', 'train_073',...
+             'train_074', 'train_075', 'train_076', 'train_077', 'train_078', 'train_079'};
+         
+    user7 = {'train_080', 'train_081', 'train_082', 'train_083', 'train_084',...
+             'train_085', 'train_086', 'train_087'};
+         
+    users_group = cat(1, {user1}, {user2}, {user3}, {user4}, {user5}, {user6}, {user7});
+    user_inds = [];
+    
+    for i=1:numel(users)
+        
+        hog_file = [hog_data_dir, '/au_training_' users{i} '.hog'];
+        
+        f = fopen(hog_file, 'r');
+                          
+        curr_data = [];
+        curr_ind = 0;
+        
+        while(~feof(f))
+                        
+            if(curr_ind == 0)
+                num_cols = fread(f, 1, 'int32');
+                if(isempty(num_cols))
+                    break;
+                end
+
+                num_rows = fread(f, 1, 'int32');
+                num_chan = fread(f, 1, 'int32');
+
+                curr_ind = curr_ind + 1;            
+
+                % preallocate some space
+                if(curr_ind == 1)
+                    curr_data = zeros(5000, 1 + num_rows * num_cols * num_chan);
+                    num_feats =  1 + num_rows * num_cols * num_chan;
+                end
+
+                if(curr_ind > size(curr_data,1))
+                    curr_data = cat(1, curr_data, zeros(6000, 1 + num_rows * num_cols * num_chan));
+                end
+                feature_vec = fread(f, [1, 1 + num_rows * num_cols * num_chan], 'float32');
+                curr_data(curr_ind, :) = feature_vec;
+            else
+            
+                % Reading in batches of 5000
+                
+                feature_vec = fread(f, [4 + num_rows * num_cols * num_chan, 5000], 'float32');
+                feature_vec = feature_vec(4:end,:)';
+                
+                num_rows_read = size(feature_vec,1);
+                
+                curr_data(curr_ind+1:curr_ind+num_rows_read,:) = feature_vec;
+                
+                curr_ind = curr_ind + size(feature_vec,1);
+                
+            end
+                        
+        end
+        
+        fclose(f);
+        
+        curr_data = curr_data(1:curr_ind,:);
+        
+        for k=1:numel(users_group)
+            if(~isempty(strmatch(users{i}, users_group{k})))
+                user_inds = cat(1, user_inds, k * ones(curr_ind,1));
+            end
+        end
+        
+        valid = logical(curr_data(:, 1));
+        
+        vid_id_curr = cell(curr_ind,1);
+        vid_id_curr(:) = users(i);
+        
+        vid_id = cat(1, vid_id, vid_id_curr);
+        
+        % Assume same number of frames per video
+        if(i==1)
+            hog_data = zeros(curr_ind*numel(users), num_feats);
+        end
+        
+        if(size(hog_data,1) < feats_filled+curr_ind)
+           hog_data = cat(1, hog_data, zeros(size(hog_data,1), num_feats));
+        end
+        
+        hog_data(feats_filled+1:feats_filled+curr_ind,:) = curr_data;
+        
+        feats_filled = feats_filled + curr_ind;
+        
+    end
+        
+    if(numel(users) > 0)        
+        
+        % Perform normalization here
+        u_id = unique(user_inds)';
+        
+        valid_data = hog_data(1:feats_filled,1) > 0;
+        
+        for u=u_id
+            hog_data(user_inds==u, 2:end) = bsxfun(@plus, hog_data(user_inds==u, 2:end), -median(hog_data(user_inds==u & valid_data, 2:end)));
+        end
+        
+        hog_data = hog_data(1:feats_filled,2:end);
+    end
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Read_geom_files.m

@@ -0,0 +1,36 @@
+function [geom_data, valid_ids] = Read_geom_files(users, params_data_dir)
+
+    geom_data = [];
+    valid_ids = [];
+    
+    load('../../pca_generation/pdm_68_aligned_wild.mat');
+    
+    for i=1:numel(users)
+        
+        geom_file = [params_data_dir, '/au_training_' users{i} '.csv'];        
+        
+        if(i == 1)
+            tab = readtable(geom_file);
+            column_names = tab.Properties.VariableNames; 
+            valid_ind = cellfun(@(x) ~isempty(x) && x==1, strfind(column_names, 'success'));
+            shape_inds = cellfun(@(x) ~isempty(x) && x==1, strfind(column_names, 'p_'));
+        end
+                
+        res = dlmread(geom_file, ',', 1, 0);    
+
+        % Check the confidence of detection
+        valid = logical(res(:, valid_ind));              
+        res = res(:, shape_inds);
+                
+        % Do not consider global parameters
+        res = res(:, 7:end);
+                          
+        actual_locs = res * V';
+        res = cat(2, actual_locs, res);
+
+        valid_ids = cat(1, valid_ids, valid);
+            
+        geom_data = cat(1, geom_data, res);
+                
+    end
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Read_geom_files_dynamic.m

@@ -0,0 +1,82 @@
+function [geom_data, valid_ids] = Read_geom_files_dynamic(users, params_data_dir)
+
+    geom_data = [];
+    valid_ids = [];
+    
+    load('../../pca_generation/pdm_68_aligned_wild.mat');   
+    
+    user1 = {'train_001', 'train_002', 'train_003', 'train_004', 'train_005',...
+            'train_006', 'train_007', 'train_008', 'train_009', 'train_010',...
+            'train_011', 'train_012', 'train_013', 'train_014', 'train_015',...
+            'train_016', 'train_017', 'train_018'};
+    
+    user2 = {'train_019', 'train_020', 'train_021', 'train_022', 'train_023',...
+             'train_024', 'train_025', 'train_026', 'train_027', 'train_028',...
+             'train_029', 'train_030', 'train_031', 'train_032'};
+        
+    user3 = {'train_033', 'train_034', 'train_035', 'train_036', 'train_037',...
+             'train_038', 'train_039', 'train_040', 'train_041'};
+         
+    user4 = {'train_042', 'train_043', 'train_044', 'train_045', 'train_046',...
+             'train_047', 'train_048', 'train_049', 'train_050', 'train_051',...
+             'train_052', 'train_053', 'train_054', 'train_055', 'train_056'};
+    
+    user5 = {'train_057', 'train_058', 'train_059', 'train_060', 'train_061',...
+             'train_062', 'train_063'};
+         
+    user6 = {'train_064', 'train_065', 'train_066', 'train_067', 'train_068',...
+             'train_069', 'train_070', 'train_071', 'train_072', 'train_073',...
+             'train_074', 'train_075', 'train_076', 'train_077', 'train_078', 'train_079'};
+         
+    user7 = {'train_080', 'train_081', 'train_082', 'train_083', 'train_084',...
+             'train_085', 'train_086', 'train_087'};
+         
+    users_group = cat(1, {user1}, {user2}, {user3}, {user4}, {user5}, {user6}, {user7});
+    user_inds = [];    
+    
+    for i=1:numel(users)
+               
+        geom_file = [params_data_dir, '/au_training_' users{i} '.csv'];        
+        
+        if(i == 1)
+            tab = readtable(geom_file);
+            column_names = tab.Properties.VariableNames; 
+            valid_ind = cellfun(@(x) ~isempty(x) && x==1, strfind(column_names, 'success'));
+            shape_inds = cellfun(@(x) ~isempty(x) && x==1, strfind(column_names, 'p_'));
+        end
+                
+        res = dlmread(geom_file, ',', 1, 0);    
+
+        % Check the confidence of detection
+        valid = logical(res(:, valid_ind));              
+        res = res(:, shape_inds);
+                
+        % Do not consider global parameters
+        res = res(:, 7:end);
+        
+        for k=1:numel(users_group)
+            if(~isempty(strmatch(users{i}, users_group{k})))
+                user_inds = cat(1, user_inds, k * ones(size(valid,1),1));
+            end
+        end
+        
+        actual_locs = res * V';
+        res = cat(2, actual_locs, res);
+
+        valid_ids = cat(1, valid_ids, valid);
+                
+
+        geom_data = cat(1, geom_data, res);
+                
+    end
+   
+    if(numel(users) > 0)        
+        
+        % Perform normalization here
+        u_id = unique(user_inds)';
+        for u=u_id
+            geom_data(user_inds==u, :) = bsxfun(@plus, geom_data(user_inds==u, :), -median(geom_data(user_inds==u & valid_ids, :)));
+        end
+    end                    
+    
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Script_HOG_SVM_train.m

@@ -0,0 +1,78 @@
+function Script_HOG_SVM_train()
+
+% Change to your downloaded location
+addpath('C:\liblinear\matlab')
+addpath('../training_code/');
+addpath('../utilities/');
+addpath('../../data extraction/');
+
+%% load shared definitions and AU data
+shared_defs;
+
+% Set up the hyperparameters to be validated
+hyperparams.c = 10.^(-9:0.5:1);
+hyperparams.e = 10.^(-3);
+
+hyperparams.validate_params = {'c', 'e'};
+
+% Set the training function
+svm_train = @svm_train_linear;
+    
+% Set the test function (the first output will be used for validation)
+svm_test = @svm_test_linear;
+
+all_recs = cat(2, train_recs, devel_recs);
+
+%%
+for a=1:numel(all_aus)
+    
+    au = all_aus(a);
+            
+    rest_aus = setdiff(all_aus, au);        
+    [train_recs, devel_recs] = get_balanced_fold(FERA2011_dir, all_recs, au, 1/3, 1);
+    
+    % load the training and testing data for the current fold    
+    [train_samples, train_labels, valid_samples, valid_labels, ~, PC, means, scaling] = Prepare_HOG_AU_data_generic(train_recs, devel_recs, au, rest_aus, FERA2011_dir, features_dir);
+    
+    train_samples = sparse(train_samples);
+    valid_samples = sparse(valid_samples);
+
+    %% Cross-validate here                
+    [ best_params, ~ ] = validate_grid_search_no_par(svm_train, svm_test, false, train_samples, train_labels, valid_samples, valid_labels, hyperparams);
+
+    model = svm_train(train_labels, train_samples, best_params);        
+
+    [~, prediction] = svm_test(valid_labels, valid_samples, model);
+
+    % Go from raw data to the prediction
+    w = model.w(1:end-1)';
+    b = model.w(end);
+
+    svs = bsxfun(@times, PC, 1./scaling') * w;
+
+    name = sprintf('models/AU_%d_static.dat', au);
+
+    pos_lbl = model.Label(1);
+    neg_lbl = model.Label(2);
+        
+    write_lin_svm(name, means, svs, b, pos_lbl, neg_lbl);
+
+    name = sprintf('results_FERA2011_devel/AU_%d_static.mat', au);
+        
+    tp = sum(valid_labels == 1 & prediction == 1);
+    fp = sum(valid_labels == 0 & prediction == 1);
+    fn = sum(valid_labels == 1 & prediction == 0);
+    tn = sum(valid_labels == 0 & prediction == 0);
+
+    precision = tp/(tp+fp);
+    recall = tp/(tp+fn);
+
+    f1 = 2 * precision * recall / (precision + recall);    
+    
+    save(name, 'model', 'f1', 'precision', 'recall', 'best_params', 'valid_labels', 'prediction');
+        
+end
+
+end
+
+

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/Script_HOG_SVM_train_dyn.m

@@ -0,0 +1,79 @@
+function Script_HOG_SVM_train_dyn()
+
+% Change to your downloaded location
+addpath('C:\liblinear\matlab')
+addpath('../training_code/');
+addpath('../utilities/');
+addpath('../../data extraction/');
+
+%% load shared definitions and AU data
+shared_defs;
+
+% Set up the hyperparameters to be validated
+hyperparams.c = 10.^(-9:0.5:1);
+hyperparams.e = 10.^(-3);
+
+hyperparams.validate_params = {'c', 'e'};
+
+% Set the training function
+svm_train = @svm_train_linear;
+    
+% Set the test function (the first output will be used for validation)
+svm_test = @svm_test_linear;
+
+all_recs = cat(2, train_recs, devel_recs);
+
+%%
+for a=1:numel(all_aus)
+    
+    au = all_aus(a);
+            
+    rest_aus = setdiff(all_aus, au);        
+    
+    [train_recs, devel_recs] = get_balanced_fold(FERA2011_dir, all_recs, au, 1/3, 1);
+    
+    % load the training and testing data for the current fold    
+    [train_samples, train_labels, valid_samples, valid_labels, ~, PC, means, scaling] = Prepare_HOG_AU_data_generic_dynamic(train_recs, devel_recs, au, rest_aus, FERA2011_dir, features_dir);
+    
+    train_samples = sparse(train_samples);
+    valid_samples = sparse(valid_samples);
+
+    %% Cross-validate here                
+    [ best_params, ~ ] = validate_grid_search_no_par(svm_train, svm_test, false, train_samples, train_labels, valid_samples, valid_labels, hyperparams);
+
+    model = svm_train(train_labels, train_samples, best_params);        
+
+    [~, prediction] = svm_test(valid_labels, valid_samples, model);
+
+    % Go from raw data to the prediction
+    w = model.w(1:end-1)';
+    b = model.w(end);
+
+    svs = bsxfun(@times, PC, 1./scaling') * w;
+
+    name = sprintf('models/AU_%d_dyn.dat', au);
+
+    pos_lbl = model.Label(1);
+    neg_lbl = model.Label(2);
+        
+    write_lin_dyn_svm(name, means, svs, b, pos_lbl, neg_lbl);
+
+    name = sprintf('results_FERA2011_devel/AU_%d_dyn.mat', au);
+
+    tp = sum(valid_labels == 1 & prediction == 1);
+    fp = sum(valid_labels == 0 & prediction == 1);
+    fn = sum(valid_labels == 1 & prediction == 0);
+    tn = sum(valid_labels == 0 & prediction == 0);
+
+    precision = tp/(tp+fp);
+    recall = tp/(tp+fn);
+
+    f1 = 2 * precision * recall / (precision + recall);    
+    
+    save(name, 'model', 'f1', 'precision', 'recall', 'best_params', 'valid_labels', 'prediction');
+        
+end
+
+end
+
+

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/get_balanced_fold.m

@@ -0,0 +1,56 @@
+function [train_users, dev_users] = get_balanced_fold(FERA_dir, users, au, prop_test, offset)
+
+
+    [ labels_train, valid_ids_train, filenames ] = extract_FERA2011_labels(FERA_dir, users, au);
+    
+    % Make sure the folds are user independent
+    user1 = {'train_001', 'train_002', 'train_003', 'train_004', 'train_005',...
+            'train_006', 'train_007', 'train_008', 'train_009', 'train_010',...
+            'train_011', 'train_012', 'train_013', 'train_014', 'train_015',...
+            'train_016', 'train_017', 'train_018'};
+    
+    user2 = {'train_019', 'train_020', 'train_021', 'train_022', 'train_023',...
+             'train_024', 'train_025', 'train_026', 'train_027', 'train_028',...
+             'train_029', 'train_030', 'train_031', 'train_032'};
+        
+    user3 = {'train_033', 'train_034', 'train_035', 'train_036', 'train_037',...
+             'train_038', 'train_039', 'train_040', 'train_041'};
+         
+    user4 = {'train_042', 'train_043', 'train_044', 'train_045', 'train_046',...
+             'train_047', 'train_048', 'train_049', 'train_050', 'train_051',...
+             'train_052', 'train_053', 'train_054', 'train_055', 'train_056'};
+    
+    user5 = {'train_057', 'train_058', 'train_059', 'train_060', 'train_061',...
+             'train_062', 'train_063'};
+         
+    user6 = {'train_064', 'train_065', 'train_066', 'train_067', 'train_068',...
+             'train_069', 'train_070', 'train_071', 'train_072', 'train_073',...
+             'train_074', 'train_075', 'train_076', 'train_077', 'train_078', 'train_079'};
+         
+    user7 = {'train_080', 'train_081', 'train_082', 'train_083', 'train_084',...
+             'train_085', 'train_086', 'train_087'};
+         
+    users_group = cat(1, {user1}, {user2}, {user3}, {user4}, {user5}, {user6}, {user7});
+    
+    counts = zeros(numel(users_group),1);
+    for k=1:numel(users_group)
+        for f=1:numel(filenames)            
+            if(~isempty(strmatch(filenames{f}(1:9), users_group{k})))
+                counts(k) = counts(k) + sum(labels_train{f});
+            end
+        end
+    end
+
+    % Now group them together by users
+    
+    [sorted, inds] = sort(counts);
+    dev_inds = inds(offset:round(1/prop_test):end);
+    train_inds = setdiff(inds, dev_inds);
+    
+    dev_users = users_group(dev_inds);
+    dev_users = cat(2, dev_users{:})';
+    
+    train_users = users_group(train_inds);
+    train_users = cat(2, train_users{:})';
+    
+end

pkg/OpenFace/model_training/AU_training/experiments/FERA2011/shared_defs.m

@@ -0,0 +1,11 @@
+% this is data defined across the experiments (to make sure all of them have same user conventions)
+
+% Defining which AU's we are extracting
+all_aus = [1, 2, 4, 6, 7, 10, 12, 15, 17, 18, 25, 26];
+
+addpath('../../data extraction/');
+
+find_FERA2011;
+
+features_dir = 'E:\datasets\face_datasets_processed\fera2011';
+