clusterCount = hyperparameters["codebook_size"](featureCount) hyperparameters["codebook_size"] = clusterCount performance["codebook_size"] = clusterCount algo.gen_codebook( TMP_DIR_TRAINING, clusterCount, currentCodebook, batch_size=algo.BATCH_SIZE if algo.BATCH_SIZE >= clusterCount else clusterCount, ) print "saved codebook to '" + currentCodebook + "'" performance["codebook"] = currentCodebook # generate histograms print "---------------------" print "## generating histograms of the training examples" algo.compute_histograms(TMP_DIR_TRAINING, currentCodebook, TMP_DIR_TRAINING) # iterate over different svm_c for svm_c in HYPERPARAMETERS_OPTIONS["svm_c"]: hyperparameters["svm_c"] = svm_c print "HYPERPARAMETER: svm_c = " + str(svm_c) # iterate over different svm_gamma for svm_gamma in HYPERPARAMETERS_OPTIONS["svm_gamma"]: hyperparameters["svm_gamma"] = svm_gamma print "HYPERPARAMETER: svm_gamma = " + str(svm_gamma) # train svm print "---------------------" print "## training svm" algo.train_svm(
certainty = float(i.replace("\\","/").rpartition("/")[2].partition("_")[0]) label = 1 if certainty > 0 else 0 all_labels[i] = label all_weights[i] = certainty if label == 1 else 1-certainty # extract features featureCount = algo.extract_features(all_files, TMP_DIR) # generate codebook clusterCount = int(sqrt(featureCount)) algo.gen_codebook(TMP_DIR, clusterCount, SIFT_CODEBOOK, batch_size = algo.BATCH_SIZE if algo.BATCH_SIZE >= clusterCount else clusterCount) # generate histograms algo.compute_histograms(TMP_DIR, SIFT_CODEBOOK, TMP_DIR) # train svm algo.train_svm(TMP_DIR, all_labels, SVM_MODEL_FILE, all_weights = all_weights) print "calculating predictions" predictions = algo.predict(SVM_MODEL_FILE, SIFT_CODEBOOK, DATASETPATH2, TMP_DIR) img = Image.open('dop' +f + '/dop-annotated.png').convert('RGBA') overlay = Image.new('RGBA', img.size, 0) draw = ImageDraw.Draw(overlay) print "\n\nPredictions:" for filepath, is_building in predictions.items():