StarGaze-BD

PART-A : STATISTICAL MODEL BASED

Description of Identifying Famous Bangladeshi Personalities Using Machine Learning Techniques

This machine learning-based web application is designed to identify famous Bangladeshi personalities based on their photographs and provide details based on that. By analyzing the visual features of the images and applying advanced algorithms, the application is able to accurately classify and recognize the individuals depicted in the photos upto a certain extent.

Steps for Adding and Preprocessing Additional Dataset for Famous Bangladeshi Personality Classification:

• Create a new folder with at least 30-40 images of a specific person, ensuring that the images clearly show their face.
• Rename the folder with the person's name (e.g. Mr Tom).
• Copy or cut and paste the folder into the following directories:
  • StarGaze-BD/model/collected_dataset
  • StarGaze-BD/model/dataset
• Run the following notebooks in the specified order:
  • StarGaze-BD/model/rotate_and_zoom_images.ipynb
  • StarGaze-BD/model/famous_bangladeshi_persona_identification_model.ipynb

To complete these instructions, you will need to have Python and the required packages installed on your system. You can install the required packages by running the command pip install -r requirements in your terminal. This will install all of the necessary packages specified in the requirements.txt file.

Technology Stack PART-A

• Machine Learning Algorithms : Random Forest, Support Vector Machine (SVM), Logistic Regression
• Machine Learning Library : scikit-learn
• Front-End : HTML, Bootstrap, JavaScript

How To Run PART-A

1. Go to Server directory
2. Open terminal on this directory
3. Type python server.py in the terminal and hit Enter
4. Go to UI directory
5. Open app.html
6. Hurrah...!!

Team Member:

• 2018331089 - Ali Al - Reza
• 2018331029 - Ishrat Jahan
• 2018331071 - Nishat Rahman
• 2018331081 - Md Ataullha
• 2018331111 - Redwanur Rahman Akanda

---

PART-B : DEEP LEARNING BASED

Description of Classifying Famous Bangladeshi Personalities Using Deep Learning CNN Technique

This deep learning-based web application is designed to identify famous Bangladeshi personalities based on their photographs and provide details based on that. By analyzing the visual features of the images and applying advanced cnn algorithm, the application is able to accurately classify and recognize the individuals depicted in the photos upto a certain extent.

Technology Stack PART-B

• Machine Learning Algorithms : Convolutional Neural Network (CNN)
• Deep Learning Framework : TensorFlow
• Web App Framework : Streamlit

How To Run PART-B

1. Go to .spyder-py3 directory
2. Open terminal on this directory
3. Type streamlit run temp.py in the terminal and hit Enter (tensorflow needed)
4. Hurrah...!!

Team Member:

• 2018331089 - Ali Al - Reza
• 2018331029 - Ishrat Jahan
• 2018331071 - Nishat Rahman
• 2018331081 - Md Ataullha
• 2018331111 - Redwanur Rahman Akanda