README.md

metadata

title: CV Lab Camera Studio
emoji: 📷
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 5.50.0
app_file: app.py
pinned: false
short_description: Computer vision, AR filters, and image analysis laboratory.
license: mit

CV Lab Camera

CV Lab Camera is a Gradio application and developer suite for computer vision, scientific image analysis, geometric transformations, morphology, neural style transfer, AR face overlays, and batch dataset processing.

---

Developer Quickstart

1. Web App

Launch the interactive Gradio laboratory:

python app.py

2. Command Line Interface (CLI)

Process images directly from your terminal or shell scripts:

# Apply built-in filter
python cli.py filter --input input.jpg --filter Sepia --output result.jpg

# Apply geometric transformation
python cli.py transform --input input.jpg --op Rotation --angle 45 --output result.jpg

# Apply morphological operation
python cli.py morph --input input.jpg --op Opening --size 5 --output result.jpg

# Apply AR face overlay
python cli.py ar --input face.jpg --filter Glasses --output ar_result.jpg

# Apply neural style transfer
python cli.py style --input content.jpg --style-image style.jpg --output stylized.jpg

# Batch process dataset
python cli.py batch --dir ./my_dataset --filter Grayscale --output-zip processed.zip

3. Python Developer SDK (developer_api.py)

Import CV Lab functions directly in your Python code:

from developer_api import process_image, generate_python_snippet

# Process image programmatically
result, meta = process_image(
    image_input="photo.png",
    operation_type="filter",
    operation_name="Sepia",
    output_path="output_sepia.png"
)

# Generate copy-pasteable Python code
code = generate_python_snippet("Sepia", {"contrast": 1.2})
print(code)

---

Architecture Overview

AI_Lab_CAM/
├── app.py                     # Gradio UI Web App (interactive tabs & code generator)
├── cli.py                     # Developer Command Line Interface (CLI)
├── developer_api.py           # Developer SDK / Python API wrapper
├── batch/
│   └── dataset_processor.py   # Batch directory/zip processing with manifest.json
├── cv_ops/
│   ├── analysis.py            # Intensity histograms & pixel statistics
│   ├── morphology.py          # Thresholding & morphological operations
│   └── transforms.py          # Translation, rotation, scaling, reflection
├── filters/
│   ├── builtin.py             # Pure NumPy / OpenCV filter functions
│   ├── custom.py              # Custom kernel & pipeline JSON parsers
│   └── registry.py            # Filter registry & saved JSON persistence
├── models/
│   ├── face_filters.py        # OpenCV AR face landmark detector & filter engine
│   └── style_transfer.py      # TensorFlow Hub Magenta neural style transfer
├── saved_filters/             # Saved custom filter JSONs
├── saved_ar_filters/          # Saved custom AR filter JSONs
├── tests/                     # Pytest suite
└── requirements.txt           # Python dependencies

---

Custom Filters & AR Filters

Image Processing Filter Pipelines

Save multi-step filter pipelines as JSON:

[
  {"operation": "Grayscale", "params": {}},
  {"operation": "Sharpen", "params": {"amount": 1.4}}
]

Custom AR Face Filters

Design landmark-based AR overlays attached to head_top, forehead, eyes, nose, mouth, chin:

{
  "elements": [
    {
      "landmark": "forehead",
      "shape": "crown",
      "color": [255, 215, 0],
      "scale": 1.0,
      "offset_y": -0.15
    },
    {
      "landmark": "eyes",
      "shape": "visor",
      "color": [0, 255, 255],
      "scale": 1.0
    }
  ]
}

---

Testing

Run the test suite:

pytest