😀 CascadeVision

This is Face Detection System which detects faces from an image or a live camera using OpenCV (Haar Cascade).

---

Project files

• Source: FaceDetection.cpp
• Demo images: Images/Avengers.jpg, Images/FF.jpg, Images/HP.jpg, Images/TC.jpg

---

🔧 What it does

• Runs face detection using OpenCV's Haar cascade (haarcascade_frontalface_alt.xml).
• Two modes:
  • Live Face Detection (camera)
  • Image Face Detection (single image)

The program draws red rectangles around detected faces and can save the output image.

---

✅ Prerequisites

• OpenCV (built for your Visual Studio version)
• Microsoft Visual Studio (any recent edition)

---

🛠 Setup (Visual Studio)

1. Install OpenCV and note the paths (example used in code):

   • Include headers: C:\OpenCV\opencv\build\include
   • Libraries: C:\OpenCV\opencv\build\x64\vc14\lib
   • Binary (optional, add to PATH): C:\OpenCV\opencv\build\x64\vc14\bin

2. Create a new Visual C++ Console project and add FaceDetection.cpp to the project.

3. In Project Properties:

   • C/C++ → General → Additional Include Directories: add the OpenCV include path above.
   • Linker → General → Additional Library Directories: add the OpenCV lib path.
   • Linker → Input → Additional Dependencies: add the appropriate opencv_world*.lib (e.g., opencv_world452.lib or opencv_world452d.lib for debug).

4. Ensure runtime DLLs are available (either add the OpenCV bin folder to PATH or copy required DLLs next to the executable).

---

▶ How to run

1. Build the project in Visual Studio.
2. Run the produced executable.
3. Choose an option at the menu:
   • 1 — Live Face Detection (requires camera)
   • 2 — Image Face Detection (the code uses a hardcoded path by default)
   • 3 — Exit

Notes on defaults in the source:

• Cascade path (loaded at startup):
  • C:\OpenCV\opencv\sources\data\haarcascades\haarcascade_frontalface_alt.xml — update this to your haarcascades location if different.
• Image mode default path (change inside imgdetect()):
  • D:\College\KIT\Projects\2nd Year\4th Sem\Mini Project\Image\tc.jpg
• Output paths (change inside livedetect() / imgdetect()):
  • D:\College\KIT\Projects\2nd Year\4th Sem\Mini Project\Output\...

Replace those hardcoded paths or modify the code to accept command-line arguments for better portability.

---

🧩 Code notes & suggestions

• The project uses CascadeClassifier::detectMultiScale for face detection.
• Hardcoded paths make the program fragile; consider:
  • Accepting command-line args or a config file for cascade, input image, and output directories.
  • Using cv::CommandLineParser or simple argv parsing.
  • Checking and creating the output directory before saving images.

Small example change (concept): Replace hardcoded cascade load with a path variable passed via argv or environment.

---

📷 Sample Results

Below are two example outputs from Images/ (click to open full-size):

---

⚠️ Known limitations

• Program exits when no face is found in image mode.
• Uses blocking infinite loops with waitKey and exit(0); for production you'd want cleaner shutdown logic.

---

🙏 Credits

• Built with OpenCV — great thanks to the OpenCV community.