Nu Metal Pose: Random Image Detector

A real-time computer vision project that detects a specific hand gesture (the "Nu Metal Pose") and displays a random image on screen.

This is an open-source Python project that uses your webcam to perform real-time gesture detection. When the application identifies two open hands on screen simultaneously, it triggers an event and overlays a random image from a local folder onto the video feed.

(Feel free to add a GIF or screenshot of your project working here!)

✨ Features

• Real-Time Hand Tracking: Utilizes MediaPipe to track 21 keypoints on multiple hands.
• Gesture Recognition: Implements custom logic to determine if a hand is "open" or "closed".
• "Nu Metal Pose" Trigger: Specifically detects when two hands are simultaneously in the "open" state.
• Random Image Overlay: When the pose is triggered, the app randomly selects one of your custom images and displays it.
• Transparency Support: Correctly overlays PNG images with alpha channels (transparency).

---

🛠️ How It Was Made (Technology)

This project was built entirely in Python using two core computer vision libraries:

• OpenCV: Used for capturing the webcam video feed, basic image processing (like flipping the video for a "selfie" mode), color space conversion (BGR to RGB), and overlaying the final images and text onto the screen.
• MediaPipe (by Google): This is the "magic" behind the gesture recognition. We use the MediaPipe Hands solution, which provides a high-fidelity, 21-point 3D landmark model for each hand it detects in the frame.

---

💡 The Logic

1. Capture Frame: The script reads a new frame from the webcam using cv2.VideoCapture().
2. Process Frame: The frame is sent to the MediaPipe Hands model for processing.
3. Find Hands: The model returns the landmark (keypoint) data for up to two hands.
4. Check Gesture: For each detected hand, a custom function is_hand_open() runs. This function checks the Y-coordinates of the fingertips (like the index finger tip, landmark #8) against their middle joints (index finger joint, landmark #6).
   • If tip.y < joint.y (meaning the tip is higher on the screen), the finger is considered "open".
   • If 4 or 5 fingers are "open", the hand is classified as open.
5. Trigger Event: If the code counts open_hands_count == 2, it sets a "pose detected" flag.
6. Pick Random Image: A state-locking mechanism (pose_detected_previously) ensures that a new random image is chosen from your folder only on the first frame the pose is detected (this prevents the image from flickering randomly).
7. Overlay Image: The chosen image is "pasted" onto the main video frame using cv2 (OpenCV), correctly handling transparency.
8. Display: The final, modified frame is shown to the user.

---

🚀 How to Run

1. Prerequisites

You must have Python 3.x installed on your machine.

2. Installation

1. Clone this repository:

   git clone [https://github.com/](https://github.com/)/GabrielaMarculino/Nu-Metal-Pose-Random-Image-Detector.git
   cd [Nome da Pasta]

2. Install the required libraries:

   pip install opencv-python mediapipe

3. Usage

1. Add Your Images:

   • Find 5 images (or more!) that you want to display.
   • Place them in the same root folder as the Python script.
   • Update the IMAGE_FILENAMES list in the Python script (main.py) with the exact names of your files:

   IMAGE_FILENAMES = [
       'my_image_1.png',  
       'my_image_2.jpg',  
       # etc...
   ]

2. Run the script:

   python main.py

   (Change main.py to whatever you named your file.)

3. Make the Pose!

   • Show both of your hands, open, to the camera to trigger the random image.

---

📄 License

This project is open source and distributed under the MIT License. This means you are free to use, modify, and distribute this code for any purpose. See the LICENSE file for more details.