Vgg-16 & Vgg-19
-Hendra
-Fatturahman
GoogLeNet
-Yogi
-Dani
ResNet50
-Harisson
-Fitrah
In the modern digital era, security and identity management have become pressing issues. One of the primary challenges faced is the efficient supervision and control of access, especially in dense and complex environments. To address this issue, Face Recognition technology has emerged as a solution.
Face Recognition is a technology that utilizes algorithms to automatically identify and verify human faces. This process involves analyzing facial features such as the shape of the nose, eyes, and mouth to create a unique template for identification. The technology has rapidly advanced and found applications in various fields, including security, facial recognition on mobile devices, and identity management.
-
High-Level Security: Face Recognition systems supported by AI provide a high level of security, identifying individuals even in challenging situations.
-
Access Management: The use of this technology in access management to restricted areas or sensitive information helps enhance control and reduce the risk of unauthorized access.
-
Crime Recognition: In law enforcement, this technology can assist in identifying and tracking criminals more efficiently.
-
Identity Management: In a business context, Face Recognition aids in employee identity management, optimizing attendance processes and reducing internal security risks.
This image depicts how Face Recognition technology collaborates with AI to identify individuals with high accuracy. The process involves analyzing facial features and machine learning, ensuring precise recognition even in less-than-ideal conditions, such as low lighting or varied facial expressions.
By adopting Face Recognition technology supported by AI, challenges in identification and security can be more effectively addressed, bringing significant benefits to various industries and identity management needs.
"We hope that the model we have created can be useful as intended.
The first thing you need to do is:
- Import libraries
Example:
from keras.preprocessing.image import ImageDataGenerator
This code for image augmentaion
- Prepare your dataset
Example:
google.colab import drivedrive.mount('/content/drive') #Connect your drive
image_path = os.listdir(os.path.join(Data_path, 'Images')) #Import your images folder from drive
list_att =(os.path.join(Data_path, 'list_attribute2.txt')) #Import your dataset from drive
df = pd.read_csv(list_att, usecols=['image_id','Male'], delimiter=r'\s+', skiprows=1) #Read your dataset, in this case i just use 2 cols for target and feature
This code for import your data using google drive, don't forget to create your own folder,
and change the Images
folder to your own folder, and the list_attribute2.txt
to your dataset
-
Split dataset into train,test,and validation
-
Preprocess your dataset
-
Build the model
-
Train your model
-
Evaluation
"The results of our model using VGG16, VGG19, GoogLeNet, and ResNet50 vary significantly.
-
This model accuracy is 94%
-
Image prediction result
-
This model accuracy is 97%
-
The Graphic Result
-
This model accuracy is 92%
-
The Confusion Matrix Result
So far, our best-performing model is GoogLeNet with high accuracy
- Basic Libraries: NumPy, Matplotlib, Pandas
- Domain-specific Libraries: OpenCV
- Deep-learning Frameworks: Keras, PyTorch, TensorFlow
That's all, for any discussion kindly contact me here: [email protected]