Lumina CV: Spatial Enhancement & Geometrical Transformations

  Live Demo: lumina-cv.streamlit.app

Lumina CV is an interactive, real-time computer vision application built in Python using Streamlit, OpenCV, and Matplotlib. It allows users to upload custom images or use a default calibration pattern, adjust processing parameters through interactive sliders, and see side-by-side comparisons of the original and processed images along with dynamic RGB/grayscale histograms and execution speeds.

---

🚀 Key Features

1. Spatial-Domain Enhancement (Point Processing)

Direct pixel-level mapping functions ($s = T(r)$) to improve image quality:

• Brightness Adjustment: Add/subtract intensity offset constant values ($s = r + \beta$).
• Contrast Adjustment: Scale intensity values using multiplication gain factors ($s = \alpha \cdot r$).
• Image Negative: Inverse intensity mapping ($s = 255 - r$) to highlight dark features.
• Log Transformation: Dynamic range compression ($s = c \cdot \ln(1 + r)$) to expand dark regions.
• Power-Law (Gamma) Transformation: Non-linear gamma correction ($s = c \cdot r^\gamma$) to adjust mid-tones.
• Thresholding: Map pixel values to binary limits (black and white) based on threshold limit $T$.
• Contrast Stretching: Linearly stretch narrow input intensity ranges to fill the full $[0, 255]$ range.

2. Geometrical Transformations

Coordinate mappings to translate, shear, scale, or warp image shapes:

• Translation: Shift image coordinates horizontally and vertically.
• Scaling: Resize image dimensions with adjustable interpolation methods (Nearest Neighbor, Bilinear, Bicubic).
• Rotation: Rotate the image by custom degrees around any pivot point (e.g. image center).
• Reflections:
  • Horizontal: Flips image horizontally.
  • Vertical: Flips image vertically.
  • Origin: Flips image both horizontally and vertically ($180^\circ$ rotation).
• Shearing: Slant the image along the X-axis or Y-axis by a shearing factor.
• Affine Transformation: Warp the image using three control points (Top-Left, Top-Right, Bottom-Left offsets) displaying the resulting $2 \times 3$ transformation matrix in real-time.

---

🛠️ Issues Encountered & Resolution

During development, the following runtime issue was encountered:

⚠️ ModuleNotFoundError: No module named 'matplotlib'

• Symptom: When launching the application (app.py), the program crashed during initialization at:

  import matplotlib.pyplot as plt

  Outputting a ModuleNotFoundError because matplotlib was missing from the virtual environment packages.
• Root Cause: Matplotlib was not included in the initial installation command, which only installed streamlit, opencv-python, numpy, and pillow.
• Resolution:
  1. Activated the virtual environment and installed the package:

     ./venv/bin/pip install matplotlib

  2. Gracefully terminated the old server process.
  3. Restarted the Streamlit server daemon:

     ./venv/bin/streamlit run app.py --server.port=8501 --server.headless=true

  4. Verified successful compilation and loading.

---

🔧 Installation & Setup

Requirements

• Python 3.10+
• pip

Step-by-step Setup

1. Clone this repository:

   git clone <repository-url>
   cd image-processor

2. Set up Virtual Environment:

   python3 -m venv venv
   source venv/bin/activate

3. Install dependencies:

   pip install streamlit opencv-python numpy pillow matplotlib

4. Run the application:

   streamlit run app.py

   The application will open automatically in your default browser at http://localhost:8501.

---

👤 Developer

Built by Dev-Aryan-Saxena (contact.aryan.dev@gmail.com).