Last active
October 13, 2021 12:11
-
-
Save prerakmody/2c12af29ef2a314a29bae32f8afd1087 to your computer and use it in GitHub Desktop.
Expected Calibration Error
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| """ | |
| Expected Calibration Error for semantic segmentation tasks | |
| """ | |
| import pdb | |
| import nrrd # pip install pynrrd | |
| import traceback | |
| import numpy as np | |
| from pathlib import Path | |
| import matplotlib.pyplot as plt | |
| FILE_DIR = Path(__file__).parent.absolute() | |
| PLOT_DIR = Path(FILE_DIR).joinpath('_tmp') | |
| Path(PLOT_DIR).mkdir(parents=True, exist_ok=True) | |
| nan_value = -0.1 | |
| def get_ece_patient(y_true, y_predict, patient_id, res_global, verbose=False, show=False): | |
| """ | |
| Params | |
| ------ | |
| y_true : [H,W,D,C], np.array, binary | |
| y_predict : [H,W,D,C], np.array, with probability values | |
| patient_id: str | |
| res_global: dict | |
| Output | |
| ------ | |
| res_global: dict | |
| Reference | |
| --------- | |
| - On Calibration of Modern Neural Networks | |
| - Non author implementation: https://github.com/sirius8050/Expected-Calibration-Error/blob/master/ECE.py | |
| """ | |
| if verbose: print ('\n - [get_ece()] patient_id: ', patient_id) | |
| try: | |
| # Step 0 - Init | |
| res = {} | |
| label_count = y_true.shape[-1] | |
| # Step 1 - Calculate outputs (in terms of label_id) | |
| o_true = np.argmax(y_true, axis=-1) | |
| o_predict = np.argmax(y_predict, axis=-1) | |
| # Step 2 - Loop over different classes | |
| for label_id in range(label_count): | |
| if verbose: print (' --- [get_ece()] label_id: ', label_id) | |
| # Step 2.1 - Make res_global for that label_id | |
| if label_id not in res_global: | |
| res_global[label_id] = {'o_predict_label':[], 'y_predict_label':[], 'o_true_label':[]} | |
| # Step 2.2 - Get o_predict_label(label_ids), o_true_label(label_ids), y_predict_label(probs) [and append to global list] | |
| o_true_label = o_true[o_predict == label_id] | |
| o_predict_label = o_predict[o_predict == label_id] | |
| y_predict_label = y_predict[:,:,:,label_id][o_predict == label_id] | |
| res_global[label_id]['o_true_label'].extend(o_true_label.flatten().tolist()) | |
| res_global[label_id]['o_predict_label'].extend(o_predict_label.flatten().tolist()) | |
| res_global[label_id]['y_predict_label'].extend(y_predict_label.flatten().tolist()) | |
| if len(o_true_label) and len(y_predict_label): | |
| # Step 2.3 - Bin the probs and calculate their mean | |
| y_predict_label_bin_ids = np.digitize(y_predict_label, np.array([0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.01]), right=False) - 1 | |
| y_predict_binned_vals = [y_predict_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| y_predict_bins_mean = [np.mean(vals) if len(vals) else nan_value for vals in y_predict_binned_vals] | |
| # Step 2.4 - Calculate the accuracy of each bin | |
| o_predict_label_bins = [o_predict_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| o_true_label_bins = [o_true_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| y_predict_bins_accuracy = [np.sum(o_predict_label_bins[bin_id] == o_true_label_bins[bin_id])/len(o_predict_label_bins[bin_id]) if len(o_predict_label_bins[bin_id]) else nan_value for bin_id in range(label_count)] | |
| y_predict_bins_len = [len(o_predict_label_bins[bin_id]) for bin_id in range(label_count)] | |
| # Step 2.5 - Wrapup | |
| N = np.prod(y_predict_label.shape) | |
| ce = np.array((np.array(y_predict_bins_len)/N)*(np.array(y_predict_bins_accuracy)-np.array(y_predict_bins_mean))) | |
| ce[ce == 0] = nan_value # i.e. y_predict_bins_accuracy[bin_id] == y_predict_bins_mean[bind_id] = nan_value | |
| res[label_id] = ce | |
| else: | |
| res[label_id] = -1 | |
| # Plot patient-wise and labelwise plots | |
| if show: | |
| print (' --- [get_ece()] y_predict_bins_accuracy: ', ['%.4f' % (each) for each in np.array(y_predict_bins_accuracy)]) | |
| print (' --- [get_ece()] CE : ', ['%.4f' % (each) for each in np.array(res[label_id])]) | |
| print (' --- [get_ece()] ECE: ', np.sum(np.abs(res[label_id][res[label_id] != nan_value]))) | |
| # GT Probs (sorted) in plt.plot (with equally-sized bins) | |
| if 0: | |
| tmp = np.sort(y_predict_label) | |
| tmp_len = len(tmp) | |
| plt.plot(range(len(tmp)), tmp, color='orange') | |
| for boundary in np.arange(0,tmp_len, int(tmp_len//10)): plt.plot([boundary, boundary], [0.0,1.0], color='black', alpha=0.5, linestyle='dashed') | |
| plt.plot([0,0],[0,0], color='black', alpha=0.5, linestyle='dashed', label='Bins(equally-sized)') | |
| plt.title('Sorted Softmax Probs (GT) (label={})\nPatient:{}'.format(label_id, patient_id)) | |
| plt.legend() | |
| # plt.show() | |
| plt.savefig(str(Path(PLOT_DIR).joinpath('ECE_SortedProbs_label_{}_{}.png'.format(label_id, patient_id))), bbox_inches='tight');plt.close() | |
| # ECE plot | |
| if 1: | |
| plt.plot(np.arange(11), np.arange(11)/10.0, linestyle='dashed', color='black', alpha=0.8) | |
| plt.scatter(np.arange(len(y_predict_bins_mean)) + 0.5 , y_predict_bins_mean, alpha=0.5, color='g', marker='s', label='Mean Pred') | |
| plt.scatter(np.arange(len(y_predict_bins_accuracy)) + 0.5 , y_predict_bins_accuracy, alpha=0.5, color='b', marker='x', label='Accuracy') | |
| diff = np.array(y_predict_bins_accuracy)-np.array(y_predict_bins_mean) | |
| for bin_id in range(len(y_predict_bins_accuracy)): plt.plot([bin_id + 0.5, bin_id + 0.5],[y_predict_bins_accuracy[bin_id], y_predict_bins_mean[bin_id]], color='pink') | |
| plt.plot([bin_id + 0.5, bin_id + 0.5],[y_predict_bins_accuracy[bin_id], y_predict_bins_mean[bin_id]], color='pink', label='CE') | |
| plt.xticks(ticks=np.arange(11), labels=np.arange(11)/10.0) | |
| plt.title('CE (label={})\nPatient:{}'.format(label_id, patient_id)) | |
| plt.xlabel('Probability') | |
| plt.ylabel('Accuracy') | |
| plt.legend() | |
| # plt.show() | |
| plt.savefig(str(Path(PLOT_DIR).joinpath('ECE_label_{}_{}.png'.format(label_id, patient_id))), bbox_inches='tight');plt.close() | |
| except: | |
| traceback.print_exc() | |
| pdb.set_trace() | |
| return res_global | |
| def get_ece_global(ece_global): | |
| try: | |
| # Step 0 - Init | |
| ece_labels_obj = {} | |
| ece_labels = [] | |
| label_count = len(ece_global) | |
| ece_global_obj_keys = list(ece_global.keys()) | |
| # Step 1 - Loop over all labelids (across all patients) | |
| for label_id in ece_global_obj_keys: | |
| o_true_label = np.array(ece_global[label_id]['o_true_label']) | |
| o_predict_label = np.array(ece_global[label_id]['o_predict_label']) | |
| y_predict_label = np.array(ece_global[label_id]['y_predict_label']) | |
| if label_id in ece_global: del ece_global[label_id] | |
| # Step 1.1 - Bin the probs and calculate their mean | |
| y_predict_label_bin_ids = np.digitize(y_predict_label, np.array([0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.01]), right=False) - 1 | |
| y_predict_binned_vals = [y_predict_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| y_predict_bins_mean = [np.mean(vals) if len(vals) else nan_value for vals in y_predict_binned_vals] | |
| # Step 1.2 - Calculate the accuracy of each bin | |
| o_predict_label_bins = [o_predict_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| o_true_label_bins = [o_true_label[y_predict_label_bin_ids == bin_id] for bin_id in range(label_count)] | |
| y_predict_bins_accuracy = [np.sum(o_predict_label_bins[bin_id] == o_true_label_bins[bin_id])/len(o_predict_label_bins[bin_id]) if len(o_predict_label_bins[bin_id]) else nan_value for bin_id in range(label_count)] | |
| y_predict_bins_len = [len(o_predict_label_bins[bin_id]) for bin_id in range(label_count)] | |
| # Step 1.3 - Wrapup | |
| N = np.prod(y_predict_label.shape) | |
| ce = np.array((np.array(y_predict_bins_len)/N)*(np.array(y_predict_bins_accuracy)-np.array(y_predict_bins_mean))) | |
| ce[ce == 0] = nan_value | |
| ece_label = np.sum(np.abs(ce[ce != nan_value])) | |
| ece_labels.append(ece_label) | |
| ece_labels_obj[label_id] = {'y_predict_bins_mean':y_predict_bins_mean, 'y_predict_bins_accuracy':y_predict_bins_accuracy, 'ce':ce, 'ece':ece_label} | |
| print ('\n') | |
| print (' - ece_labels : ', ['%.4f' % each for each in ece_labels]) | |
| print (' - ece : %.4f' % np.mean(ece_labels)) | |
| print (' - ece (w/o bgd): %.4f' % np.mean(ece_labels[1:])) | |
| print (' - ece (w/o bgd, w/o chiasm): %.4f' % np.mean(ece_labels[1:2] + ece_labels[3:])) | |
| # Step 2 - Plot ECE | |
| for label_id in ece_labels_obj: | |
| y_predict_bins_mean = ece_labels_obj[label_id]['y_predict_bins_mean'] | |
| y_predict_bins_accuracy = ece_labels_obj[label_id]['y_predict_bins_accuracy'] | |
| ece = ece_labels_obj[label_id]['ece'] | |
| plt.plot(np.arange(11), np.arange(11)/10.0, linestyle='dashed', color='black', alpha=0.8) | |
| plt.scatter(np.arange(len(y_predict_bins_mean)) + 0.5 , y_predict_bins_mean, alpha=0.5, color='g', marker='s', label='Mean Pred') | |
| plt.scatter(np.arange(len(y_predict_bins_accuracy)) + 0.5 , y_predict_bins_accuracy, alpha=0.5, color='b', marker='x', label='Accuracy') | |
| for bin_id in range(len(y_predict_bins_accuracy)): plt.plot([bin_id + 0.5, bin_id + 0.5],[y_predict_bins_accuracy[bin_id], y_predict_bins_mean[bin_id]], color='pink') | |
| plt.plot([bin_id + 0.5, bin_id + 0.5],[y_predict_bins_accuracy[bin_id], y_predict_bins_mean[bin_id]], color='pink', label='CE') | |
| plt.xticks(ticks=np.arange(11), labels=np.arange(11)/10.0) | |
| plt.title('CE (label={})\nECE: {}'.format(label_id, '%.5f' % (ece))) | |
| plt.xlabel('Probability') | |
| plt.ylabel('Accuracy') | |
| plt.ylim([-0.15, 1.05]) | |
| plt.legend() | |
| # plt.show() | |
| path_results = str(Path(PLOT_DIR).joinpath('results_ece_label{}.png'.format(label_id))) | |
| plt.savefig(str(path_results), bbox_inches='tight') | |
| plt.close() | |
| except: | |
| traceback.print_exc() | |
| pdb.set_trace() | |
| if __name__ == "__main__": | |
| patient_ids = ['p0522c0667', 'p0522c0661'] | |
| ece_global = {} | |
| for patient_id in patient_ids: | |
| y_true, _ = nrrd.read(str(Path(PLOT_DIR).joinpath('{}_true.nrrd'.format(patient_id)))) | |
| y_predict, _ = nrrd.read(str(Path(PLOT_DIR).joinpath('{}_predict.nrrd'.format(patient_id)))) | |
| patient_id = patient_id | |
| ece_global = get_ece_patient(y_true, y_predict, patient_id, ece_global, verbose=True, show=False) | |
| get_ece_global(ece_global) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi Mody, as we discussed, your "for bin_id in range(label_count)" needs to be replaced by "for bin_id in range(bin_count)" and put the following two lines at very first of the file.