import sys sys.path.append("..") import pandas as pd import numpy as np import os import re import matplotlib.pyplot as plt import scipy.stats as ss import scikit_posthocs as sp import pandas as pd from nonparametric_tests import friedman_aligned_ranks_test as ft import Orange def sort_nicely(l): """ Sort the given list in the way that humans expect. """ convert = lambda text: int(text) if text.isdigit() else text alphanum_key = lambda key: [ convert(c) for c in re.split('([0-9]+)', key) ] l.sort( key=alphanum_key ) return l #Dataset on which perform analysis #DATASET = 'LGI-PPGI' #DATASET = 'PURE' DATASET = 'UBFC1' #DATASET = 'UBFC2' #DATASET = 'Cohface' #DATASET = 'Mahnob' #DATASET = 'UBFC_ALL' CASE = 'full' #CASE = 'split' alpha = '0.05' if DATASET == 'UBFC_ALL': exp_path1 = '../../results/' + 'UBFC1' + '/' files1 = sort_nicely(os.listdir(exp_path1)) exp_path2 = '../../results/' + 'UBFC2' + '/' files2 = sort_nicely(os.listdir(exp_path2)) else: #Experiment Path exp_path = '../../results/' + DATASET + '/' files = sort_nicely(os.listdir(exp_path)) #All rPPG methods used all_methods = ['CHROM','Green','ICA','LGI','PBV','PCA','POS','SSR'] #Method(s) for the visualization of the performance vs winSize #methods = ['POS', 'CHROM', 'LGI'] #Metrics to Visualize #metrics = ['CC', 'MAE', 'RMSE'] metrics = ['MAE'] print(all_methods) #---------------- Produce Box plots for each method on a given dataset ----------- win_to_use = 10 if DATASET == 'UBFC_ALL': f_to_use = [i for i in files1 if 'winSize'+str(win_to_use) in i][0] path = exp_path1 + f_to_use res1 = pd.read_hdf(path) f_to_use = [i for i in files2 if 'winSize'+str(win_to_use) in i][0] path = exp_path2 + f_to_use res2 = pd.read_hdf(path) res = res1.append(res2) else: f_to_use = [i for i in files if 'winSize'+str(win_to_use) in i][0] path = exp_path + f_to_use res = pd.read_hdf(path) print('\n\n\t\t' + DATASET + '\n\n') if DATASET == 'UBFC1' or DATASET == 'UBFC2' or DATASET == 'Mahnob' or DATASET == 'UBFC_ALL' or DATASET == 'Cohface': all_vals_CC = [] all_vals_MAE = [] all_vals_RMSE = [] for metric in metrics: for method in all_methods: #print(method) mean_v = [] raw_values = res[res['method'] == method][metric] print(raw_values) values = [] for v in raw_values: if metric == 'CC': values.append(v[np.argmax(v)]) else: values.append(v[np.argmin(v)]) if metric == 'CC': all_vals_CC.append(np.array(values)) if metric == 'MAE': all_vals_MAE.append(np.array(values)) data_MAE = np.zeros([len(all_vals_MAE[0]), len(all_vals_MAE)]) for i,m in enumerate(all_vals_MAE): data_MAE[:,i] = m print(data_MAE) '''data_MAE_df = pd.DataFrame(data_MAE, columns=all_methods) print('\nFriedman Test MAE:') print(ss.friedmanchisquare(*data_MAE.T)) print(' ')''' '''pc = sp.posthoc_nemenyi_friedman(data_MAE_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test MAE')''' n_datasets = data_MAE.shape[0] t,p,ranks_mae,piv_mae = ft(data_MAE[:,0], data_MAE[:,1], data_MAE[:,2], data_MAE[:,3], data_MAE[:,4], data_MAE[:,5], data_MAE[:,6], data_MAE[:,7]) avranksMAE = list(np.divide(ranks_mae, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') data_CC = np.zeros([len(all_vals_CC[0]), len(all_vals_CC)]) for i,m in enumerate(all_vals_CC): data_CC[:,i] = m '''data_CC_df = pd.DataFrame(data_CC, columns=all_methods) print('\nFriedman Test MAE:') print(ss.friedmanchisquare(*data_CC.T)) print(' ') pc = sp.posthoc_nemenyi_friedman(data_CC_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test CC')''' t,p,ranks_cc,piv_cc = ft(data_CC[:,0], data_CC[:,1], data_CC[:,2], data_CC[:,3], data_CC[:,4], data_CC[:,5], data_CC[:,6], data_CC[:,7]) avranksCC = list(np.divide(ranks_cc, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') #plt.figure() #plt.subplot(1,2,1) #plt.title('CC') #plt.boxplot(all_vals_CC, showfliers=False) #plt.xticks(np.arange(1,len(all_methods)+1), all_methods) #plt.subplot(1,2,2) #plt.title('MAE') #plt.boxplot(all_vals_MAE, showfliers=False) #plt.xticks(np.arange(1,len(all_methods)+1), all_methods) cd = Orange.evaluation.compute_CD(avranksMAE, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksMAE, all_methods, cd=cd, width=6, textspace=1.5, reverse=True) #plt.title('CD Diagram MAE') cd = Orange.evaluation.compute_CD(avranksCC, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksCC, all_methods, cd=cd, width=6, textspace=1.5) #plt.title('CD Diagram CC') #plt.show() elif DATASET == 'PURE': cases = {'01':'steady', '02':'talking', '03':'slow_trans', '04':'fast_trans', '05':'small_rot', '06':'fast_rot'} all_CC = {'01':[], '02':[], '03':[], '04':[], '05':[], '06':[]} all_MAE = {'01':[], '02':[], '03':[], '04':[], '05':[], '06':[]} if CASE == 'split': for metric in metrics: for method in all_methods: #print(method) for curr_case in cases.keys(): curr_res = res[res['videoName'].str.split('/').str[5].str.split('-').str[1] == curr_case] raw_values = curr_res[curr_res['method'] == method][metric] values = [] for v in raw_values: if metric == 'CC': values.append(v[np.argmax(v)]) else: values.append(v[np.argmin(v)]) if metric == 'CC': all_CC[curr_case].append(np.array(values)) if metric == 'MAE': all_MAE[curr_case].append(np.array(values)) for curr_case in cases.keys(): '''plt.figure() plt.subplot(1,2,1) plt.title('CC ' + cases[curr_case]) plt.boxplot(all_CC[curr_case], showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) plt.subplot(1,2,2) plt.title('MAE ' + cases[curr_case]) plt.boxplot(all_MAE[curr_case], showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods)''' print('\n' + curr_case + '\n') data_MAE = np.zeros([len(all_MAE[curr_case][0]), len(all_MAE[curr_case])]) for i,m in enumerate(all_MAE[curr_case]): data_MAE[:,i] = m n_datasets = data_MAE.shape[0] data_CC = np.zeros([len(all_CC[curr_case][0]), len(all_CC[curr_case])]) for i,m in enumerate(all_CC[curr_case]): data_CC[:,i] = m t,p,ranks_mae,piv_mae = ft(data_MAE[:,0], data_MAE[:,1], data_MAE[:,2], data_MAE[:,3], data_MAE[:,4], data_MAE[:,5], data_MAE[:,6], data_MAE[:,7]) avranksMAE = list(np.divide(ranks_mae, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') t,p,ranks_cc,piv_cc = ft(data_CC[:,0], data_CC[:,1], data_CC[:,2], data_CC[:,3], data_CC[:,4], data_CC[:,5], data_CC[:,6], data_CC[:,7]) avranksCC = list(np.divide(ranks_cc, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') cd = Orange.evaluation.compute_CD(avranksMAE, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksMAE, all_methods, cd=cd, width=6, textspace=1.5, reverse=True) plt.title('CD Diagram MAE') cd = Orange.evaluation.compute_CD(avranksCC, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksCC, all_methods, cd=cd, width=6, textspace=1.5) plt.title('CD Diagram CC') plt.show() elif CASE == 'full': CC_allcases = [] MAE_allcases = [] for metric in metrics: for method in all_methods: raw_values = res[res['method'] == method][metric] values = [] for v in raw_values: if metric == 'CC': values.append(v[np.argmax(v)]) else: values.append(v[np.argmin(v)]) if metric == 'CC': CC_allcases.append(np.array(values)) if metric == 'MAE': MAE_allcases.append(np.array(values)) data_MAE = np.zeros([len(MAE_allcases[0]), len(MAE_allcases)]) for i,m in enumerate(MAE_allcases): data_MAE[:,i] = m n_datasets = data_MAE.shape[0] '''data_MAE_df = pd.DataFrame(data_MAE, columns=all_methods) print('\nFriedman Test MAE:') print(ss.friedmanchisquare(*data_MAE.T)) print(' ') pc = sp.posthoc_nemenyi_friedman(data_MAE_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test MAE')''' t,p,ranks_mae,piv_mae = ft(data_MAE[:,0], data_MAE[:,1], data_MAE[:,2], data_MAE[:,3], data_MAE[:,4], data_MAE[:,5], data_MAE[:,6], data_MAE[:,7]) avranksMAE = list(np.divide(ranks_mae, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') data_CC = np.zeros([len(CC_allcases[0]), len(CC_allcases)]) for i,m in enumerate(CC_allcases): data_CC[:,i] = m '''data_CC_df = pd.DataFrame(data_CC, columns=all_methods) print('\nFriedman Test MAE:') print(ss.friedmanchisquare(*data_CC.T)) print(' ') pc = sp.posthoc_nemenyi_friedman(data_CC_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test CC')''' t,p,ranks_cc,piv_cc = ft(data_CC[:,0], data_CC[:,1], data_CC[:,2], data_CC[:,3], data_CC[:,4], data_CC[:,5], data_CC[:,6], data_CC[:,7]) avranksCC = list(np.divide(ranks_cc, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') '''plt.figure() plt.subplot(1,2,1) plt.title('CC') plt.boxplot(CC_allcases, showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) plt.subplot(1,2,2) plt.title('MAE') plt.boxplot(MAE_allcases, showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods)''' cd = Orange.evaluation.compute_CD(avranksMAE, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksMAE, all_methods, cd=cd, width=6, textspace=1.5, reverse=True) plt.title('CD Diagram MAE') cd = Orange.evaluation.compute_CD(avranksCC, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksCC, all_methods, cd=cd, width=6, textspace=1.5) plt.title('CD Diagram CC') plt.show() elif DATASET == 'LGI-PPGI': cases = ['gym', 'resting', 'rotation', 'talk'] all_CC = {'gym':[], 'resting':[], 'rotation':[], 'talk':[]} all_MAE = {'gym':[], 'resting':[], 'rotation':[], 'talk':[]} if CASE == 'split': for metric in metrics: for method in all_methods: #print(method) for curr_case in cases: curr_res = res[res['videoName'].str.split('/').str[6].str.split('_').str[1] == curr_case] raw_values = curr_res[curr_res['method'] == method][metric] values = [] for v in raw_values: if metric == 'CC': values.append(v[np.argmax(v)]) else: values.append(v[np.argmin(v)]) if metric == 'CC': all_CC[curr_case].append(np.array(values)) if metric == 'MAE': all_MAE[curr_case].append(np.array(values)) for curr_case in cases: plt.figure() plt.subplot(1,2,1) plt.title('CC ' + curr_case) plt.boxplot(all_CC[curr_case], showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) plt.subplot(1,2,2) plt.title('MAE ' + curr_case) plt.boxplot(all_MAE[curr_case], showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) print('\n' + curr_case + '\n') data_MAE = np.zeros([len(all_MAE[curr_case][0]), len(all_MAE[curr_case])]) for i,m in enumerate(all_MAE[curr_case]): data_MAE[:,i] = m n_datasets = data_MAE.shape[0] data_CC = np.zeros([len(all_CC[curr_case][0]), len(all_CC[curr_case])]) for i,m in enumerate(all_CC[curr_case]): data_CC[:,i] = m t,p,ranks_mae,piv_mae = ft(data_MAE[:,0], data_MAE[:,1], data_MAE[:,2], data_MAE[:,3], data_MAE[:,4], data_MAE[:,5], data_MAE[:,6], data_MAE[:,7]) avranksMAE = list(np.divide(ranks_mae, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') t,p,ranks_cc,piv_cc = ft(data_CC[:,0], data_CC[:,1], data_CC[:,2], data_CC[:,3], data_CC[:,4], data_CC[:,5], data_CC[:,6], data_CC[:,7]) avranksCC = list(np.divide(ranks_cc, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') cd = Orange.evaluation.compute_CD(avranksMAE, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksMAE, all_methods, cd=cd, width=6, textspace=1.5, reverse=True) plt.title('CD Diagram MAE') cd = Orange.evaluation.compute_CD(avranksCC, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksCC, all_methods, cd=cd, width=6, textspace=1.5) plt.title('CD Diagram CC') plt.show() elif CASE == 'full': CC_allcases = [] MAE_allcases = [] for metric in metrics: for method in all_methods: raw_values = res[res['method'] == method][metric] values = [] for v in raw_values: if metric == 'CC': values.append(v[np.argmax(v)]) else: values.append(v[np.argmin(v)]) if metric == 'CC': CC_allcases.append(np.array(values)) if metric == 'MAE': MAE_allcases.append(np.array(values)) data_MAE = np.zeros([len(MAE_allcases[0]), len(MAE_allcases)]) for i,m in enumerate(MAE_allcases): data_MAE[:,i] = m n_datasets = data_MAE.shape[0] data_MAE_df = pd.DataFrame(data_MAE, columns=all_methods) print('\nFriedman Test MAE:') print(ss.friedmanchisquare(*data_MAE.T)) print(' ') pc = sp.posthoc_nemenyi_friedman(data_MAE_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test MAE') t,p,ranks_mae,piv_mae = ft(data_MAE[:,0], data_MAE[:,1], data_MAE[:,2], data_MAE[:,3], data_MAE[:,4], data_MAE[:,5], data_MAE[:,6], data_MAE[:,7]) avranksMAE = list(np.divide(ranks_mae, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') data_CC = np.zeros([len(CC_allcases[0]), len(CC_allcases)]) for i,m in enumerate(CC_allcases): data_CC[:,i] = m data_CC_df = pd.DataFrame(data_CC, columns=all_methods) print('\nFriedman Test CC:') print(ss.friedmanchisquare(*data_CC.T)) print(' ') pc = sp.posthoc_nemenyi_friedman(data_CC_df) cmap = ['1', '#fb6a4a', '#08306b', '#4292c6', '#c6dbef'] heatmap_args = {'cmap': cmap, 'linewidths': 0.25, 'linecolor': '0.5', 'clip_on': False, 'square': True, 'cbar_ax_bbox': [0.80, 0.35, 0.04, 0.3]} plt.figure() sp.sign_plot(pc, **heatmap_args) plt.title('Nemenyi Test CC') t,p,ranks_cc,piv_cc = ft(data_CC[:,0], data_CC[:,1], data_CC[:,2], data_CC[:,3], data_CC[:,4], data_CC[:,5], data_CC[:,6], data_CC[:,7]) avranksCC = list(np.divide(ranks_cc, n_datasets)) print('statistic: ' + str(t)) print('pvalue: ' + str(p)) print(' ') plt.figure() plt.subplot(1,2,1) plt.title('CC') plt.boxplot(CC_allcases, showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) plt.subplot(1,2,2) plt.title('MAE') plt.boxplot(MAE_allcases, showfliers=False) plt.xticks(np.arange(1,len(all_methods)+1), all_methods) cd = Orange.evaluation.compute_CD(avranksMAE, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksMAE, all_methods, cd=cd, width=6, textspace=1.5, reverse=True) plt.title('CD Diagram MAE') cd = Orange.evaluation.compute_CD(avranksCC, n_datasets, alpha=alpha) #tested on 30 datasets Orange.evaluation.graph_ranks(avranksCC, all_methods, cd=cd, width=6, textspace=1.5) plt.title('CD Diagram CC') plt.show()