Dark Pattern in ecommerce website and mobile applications

Implemented dark pattern categorization with YOLOv8 and uploaded results to Firebase for verification and tracking evolution.

Dark Pattern Categorization using YOLOv8

The goal of this project is to automate the identification and categorization of dark patterns in e-commerce websites and mobile applications. These patterns are manipulative tactics used to trick users into making unintended decisions.

Working Process

1. Input: The user needs to provide a screen recording of the e-commerce website or app they wish to analyze. This recording should be saved in the video/ directory.
2. Frame Extraction: The screen recording is divided into individual frames, which allows the YOLOv8 model to process each frame separately, file is not presented then the code creates by its own.
3. Detection: The frames are fed into the YOLOv8 object detection model. This model has been trained on a custom dataset that contains various types of dark patterns.
4. Output: The YOLOv8 model outputs images with bounding boxes around the detected dark patterns. Each bounding box is labeled with the class of the dark pattern.
5. Upload: The identified dark pattern images, along with their class labels, are uploaded to Firebase for further verification and to track their evolution. Additionally, a CSV file containing the image ID and identified class labels is generated and uploaded.

Dataset Creation

• The dataset used to train the YOLOv8 model was created manually using Roboflow.
• Images were sourced from various e-commerce websites.
• Each image in the dataset was manually annotated to accurately label the dark patterns.

Files and Directories

• trained_models/: Contains the YOLOv8 model files, including weights and configuration files.
• main: Contains the Python scripts for frame extraction, object detection, and uploading results to Firebase.
• output/: Stores the output images with bounding boxes and class labels.
• database.csv: A CSV file that lists the image IDs and the identified class labels for each detected dark pattern.

Installation Instructions

1. Clone the repository to your local machine.
2. Install the required Python packages.

Usage Instructions

1. Place your screen recording file in the video/ directory.
2. Run the main.py script with the appropriate input and output arguments.
3. The script will process the video, detect dark patterns, and save the results in the output/ directory. The results will also be uploaded to Firebase.

Contributing

Contributions to this project are welcome. If you have improvements or new features to add, please fork the repository and submit a pull request. Excepting for User interface.