Open-Source-SW_Term-Project

Introduction

This project utilizes OpenCV and the Caffe framework to detect faces in real-time camera input and predict the gender and age of detected faces.
It is designed for students interested in computer vision and deep learning, providing a foundational structure and code for learning and experimentation.

---

Usage

1. Run the Main Script After downloading the required models and ensuring the directory structure matches the setup provided, no additional configuration is needed. Simply execute the following command : python Gender_age_detector.py

2. The program will activate the webcam, detect faces, and display age and gender predictions in real-time.

---

Result

---

Features

• Real-time Face Detection: Detect faces from live camera input using OpenCV.
• Gender Prediction: Predict the gender of detected faces using a pre-trained Caffe model.
• Age Prediction: Estimate the age of detected faces using a pre-trained Caffe model.

---

File List

1. Gender_age_detector.py

   • The main execution file for the project.
   • Key functions:
     • Processes real-time camera input using OpenCV.
     • Detects faces and predicts their gender and age using Caffe models.
     • Displays the results in real-time on the screen.
   • Main libraries used: OpenCV, NumPy.

2. age_train_val.prototxt

   • Defines the training and validation network structure for the age prediction model in Caffe.
   • Key contents:
     • Specifies input data, network layers, and output data structure.
     • Includes hyperparameters and settings for training.

3. deploy_age.prototxt

   • Defines the deployment network structure for the age prediction model in Caffe.
   • Key contents:
     • A lightweight network structure used for testing or deploying the trained model.

4. gender_train_val.prototxt

   • Defines the training and validation network structure for the gender prediction model in Caffe.
   • Key contents:
     • Similar to the age prediction file but optimized for gender classification.

5. deploy_gender.prototxt

   • Defines the deployment network structure for the gender prediction model in Caffe.
   • Key contents:
     • A simplified network structure used for testing or deploying the trained gender prediction model.

6. style.xml

   • GUI style configuration file.
   • Key contents:
     • Specifies visual styles such as bounding box colors and text fonts for displaying results.

7. Model Files (Download Required)

   • age_net.caffemodel: Pre-trained model for age prediction (download required).
   • gender_net.caffemodel: Pre-trained model for gender prediction (download required).
   • Download locations are provided in the README or project source comments.

8. data/ (Expected Directory)

   • Directory to store datasets for model training or testing.
   • May require specifying the path in the project.

9. models/ (Expected Directory)

   • Directory to store downloaded pre-trained model files.
   • e.g., age_net.caffemodel, gender_net.caffemodel.

10. scripts/ (Expected Directory)

    • Scripts for data preprocessing or model testing.
    • e.g., dataset transformation, result visualization scripts.

---

Sources

1. Link: Caffe Model Zoo
   Description: A repository of various deep learning models for experimentation with the Caffe framework.
   Original Author: BVLC (Berkeley Vision and Learning Center)

2. Link: Age and Gender Classification
   Description: CNN-based age and gender classification models and resources created by Gil Levi.
   Original Author: Gil Levi