github/deepface
1
1
Fork 0
mirror of https://github.com/serengil/deepface.git synced 2025-06-07 12:05:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

ensemble bug

Browse Source
This commit is contained in:
Sefik Ilkin Serengil 2022-03-24 20:40:43 +00:00
parent d080107998
commit 84a6446040
1 changed files with 30 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
tests/Ensemble-Face-Recognition.py
View File

@ -25,7 +25,6 @@ idendities = {
"Matt": ["img29.jpg", "img30.jpg", "img31.jpg", "img32.jpg", "img33.jpg"], "Matt": ["img29.jpg", "img30.jpg", "img31.jpg", "img32.jpg", "img33.jpg"],
"Leonardo": ["img34.jpg", "img35.jpg", "img36.jpg", "img37.jpg"], "Leonardo": ["img34.jpg", "img35.jpg", "img36.jpg", "img37.jpg"],
"George": ["img38.jpg", "img39.jpg", "img40.jpg", "img41.jpg"] "George": ["img38.jpg", "img39.jpg", "img40.jpg", "img41.jpg"]
} }
#-------------------------- #--------------------------
#Positives #Positives
@ -33,7 +32,7 @@ idendities = {
positives = [] positives = []
for key, values in idendities.items(): for key, values in idendities.items():
#print(key) #print(key)
for i in range(0, len(values)-1): for i in range(0, len(values)-1):
for j in range(i+1, len(values)): for j in range(i+1, len(values)):
@ -45,6 +44,7 @@ for key, values in idendities.items():
positives = pd.DataFrame(positives, columns = ["file_x", "file_y"]) positives = pd.DataFrame(positives, columns = ["file_x", "file_y"])
positives["decision"] = "Yes" positives["decision"] = "Yes"
print(positives.shape) print(positives.shape)
#-------------------------- #--------------------------
#Negatives #Negatives
@ -55,18 +55,18 @@ negatives = []
for i in range(0, len(idendities) - 1): for i in range(0, len(idendities) - 1):
for j in range(i+1, len(idendities)): for j in range(i+1, len(idendities)):
#print(samples_list[i], " vs ",samples_list[j]) #print(samples_list[i], " vs ",samples_list[j])
cross_product = itertools.product(samples_list[i], samples_list[j]) cross_product = itertools.product(samples_list[i], samples_list[j])
cross_product = list(cross_product) cross_product = list(cross_product)
#print(cross_product) #print(cross_product)
for cross_sample in cross_product: for cross_sample in cross_product:
#print(cross_sample[0], " vs ", cross_sample[1]) #print(cross_sample[0], " vs ", cross_sample[1])
negative = [] negative = []
negative.append(cross_sample[0]) negative.append(cross_sample[0])
negative.append(cross_sample[1]) negative.append(cross_sample[1])
negatives.append(negative) negatives.append(negative)
negatives = pd.DataFrame(negatives, columns = ["file_x", "file_y"]) negatives = pd.DataFrame(negatives, columns = ["file_x", "file_y"])
negatives["decision"] = "No" negatives["decision"] = "No"
@ -80,8 +80,11 @@ df = pd.concat([positives, negatives]).reset_index(drop = True)
print(df.decision.value_counts()) print(df.decision.value_counts())
df.file_x = "deepface/tests/dataset/"+df.file_x df.file_x = "dataset/"+df.file_x
df.file_y = "deepface/tests/dataset/"+df.file_y df.file_y = "dataset/"+df.file_y
print(df.head())
#-------------------------- #--------------------------
#DeepFace #DeepFace
@ -94,7 +97,7 @@ pretrained_models["VGG-Face"] = VGGFace.loadModel()
print("VGG-Face loaded") print("VGG-Face loaded")
pretrained_models["Facenet"] = Facenet.loadModel() pretrained_models["Facenet"] = Facenet.loadModel()
print("Facenet loaded") print("Facenet loaded")
pretrained_models["OpenFace"] = OpenFace.loadModel() pretrained_models["OpenFace"] = OpenFace.loadModel()
print("OpenFace loaded") print("OpenFace loaded")
pretrained_models["DeepFace"] = FbDeepFace.loadModel() pretrained_models["DeepFace"] = FbDeepFace.loadModel()
print("FbDeepFace loaded") print("FbDeepFace loaded")
@ -111,7 +114,8 @@ if True:
resp_obj = DeepFace.verify(instances resp_obj = DeepFace.verify(instances
, model_name = model , model_name = model
, model = pretrained_models[model] , model = pretrained_models[model]
, distance_metric = metric) , distance_metric = metric
, enforce_detection = False)
distances = [] distances = []
@ -120,7 +124,7 @@ if True:
distances.append(distance) distances.append(distance)
df['%s_%s' % (model, metric)] = distances df['%s_%s' % (model, metric)] = distances
df.to_csv("face-recognition-pivot.csv", index = False) df.to_csv("face-recognition-pivot.csv", index = False)
else: else:
df = pd.read_csv("face-recognition-pivot.csv") df = pd.read_csv("face-recognition-pivot.csv")
@ -135,14 +139,14 @@ fig = plt.figure(figsize=(15, 15))
figure_idx = 1 figure_idx = 1
for model in models: for model in models:
for metric in metrics: for metric in metrics:
feature = '%s_%s' % (model, metric) feature = '%s_%s' % (model, metric)
ax1 = fig.add_subplot(4, 2, figure_idx) ax1 = fig.add_subplot(len(models) * len(metrics), len(metrics), figure_idx)
df[df.decision == "Yes"][feature].plot(kind='kde', title = feature, label = 'Yes', legend = True) df[df.decision == "Yes"][feature].plot(kind='kde', title = feature, label = 'Yes', legend = True)
df[df.decision == "No"][feature].plot(kind='kde', title = feature, label = 'No', legend = True) df[df.decision == "No"][feature].plot(kind='kde', title = feature, label = 'No', legend = True)
figure_idx = figure_idx + 1 figure_idx = figure_idx + 1
plt.show() plt.show()
@ -204,6 +208,15 @@ gbm.save_model("face-recognition-ensemble-model.txt")
predictions = gbm.predict(x_test) predictions = gbm.predict(x_test)
prediction_classes = []
for prediction in predictions:
prediction_class = np.argmax(prediction)
prediction_classes.append(prediction_class)
print(y_test)
print("------")
print(prediction_classes)
cm = confusion_matrix(y_test, prediction_classes) cm = confusion_matrix(y_test, prediction_classes)
print(cm) print(cm)
@ -232,7 +245,7 @@ plt.rcParams["figure.figsize"] = [20, 20]
for i in range(0, gbm.num_trees()): for i in range(0, gbm.num_trees()):
ax = lgb.plot_tree(gbm, tree_index = i) ax = lgb.plot_tree(gbm, tree_index = i)
plt.show() plt.show()
if i == 2: if i == 2:
break break
#-------------------------- #--------------------------
@ -246,4 +259,4 @@ auc = metrics.roc_auc_score(y_test, y_pred_proba)
plt.figure(figsize=(7,3)) plt.figure(figsize=(7,3))
plt.plot(fpr,tpr,label="data 1, auc="+str(auc)) plt.plot(fpr,tpr,label="data 1, auc="+str(auc))
#-------------------------- #--------------------------