Microorganism Computer Vision 🧫

A comprehensive computer vision pipeline for analyzing microorganism 🦠 motility patterns and diffusion characteristics.

Deadline: 28th February 2025

🌟 Features

• Multi-modal Tracking
  • Optical flow-based motion estimation
  • Kalman-filter enhanced object detection
  • Synthetic trajectory generation
• Advanced Diffusion Analysis
  • 6 statistical hypothesis tests for motility patterns
  • Automated trajectory classification
  • Comprehensive visualization toolkit
• Scalable Architecture
  • Modular pipeline design
  • Configurable hypothesis parameters
  • Batch processing capabilities

🧩 Diffusion Hypothesis Tests

Our framework implements sophisticated statistical tests to characterize microbial motion:

Hypothesis Test	Key Mechanism	Biological Relevance
Circular Motion CircularMotionTest	Curvature analysis + rotation angle detection	Magnetotactic bacteria, spiral-seeking organisms
Directional Switching DirectionalSwitchTest	Fourier analysis of angular changes	Run-and-tumble motility (e.g., E. coli)
Gaussian Random Walk GaussianRandomWalkTest	Kolmogorov-Smirnov normality test	Passive diffusion in isotropic environments
MSD Linear Fit MSDLinearTest	R² evaluation of mean squared displacement	Normal diffusion processes
Persistent Motion PersistentMotionTest	Velocity autocorrelation analysis	Active transport mechanisms
Subdiffusion SubDiffusionTest	Power law exponent ($\alpha < 1$) detection	Crowded environments, viscoelastic media

🚀 Installation

# Clone repository
git clone [email protected]:Ophiase/Microorganism-Computer-Vision.git
cd Microorganism-Computer-Vision

# Install dependencies
make pip

# Download sample dataset
make extract

🧪 Usage

Process real videos:

# Full pipeline execution
make transform      # Preprocess videos
make detection      # Track microorganisms
make render         # Generate gifs
make analysis       # Run statistical tests and graphics

Generate synthetic data:

make synthetic      # Create trajectory datasets
python3 -m script.main --task analysis --video synthetic_brownian
# "brownian" can be replaced by directed, sinusoidal, confined, ctrw, ...

Key Makefile Targets:

detection     # Track objects in videos
render        # Generate trajectory visualizations
analysis      # Perform statistical testing
synthetic     # Generate synthetic trajectories

📊 Results Interpretation

• 

• 

📂 Data Structure

microorganism-cv/
├── data/                           # Raw and processed data
│   ├── preprocessed/               # Normalized video tensors
│   └── tracking/                   # Trajectory datasets
├── logic/                          # Core algorithms
│   ├── diffusion/                  # Statistical tests
│   ├── filters/                    # Image processing
│   └── structure/                  # Data models
├── script/                         # Pipeline components
└── visualization/                  # Plotting utilities

📚 Methodology

Our pipeline implements a multi-stage analysis process:

Preprocessing

• Optical flow estimation
• Spatial normalization
• Noise reduction

Object Detection

• Adaptive thresholding (❌ not implemented)
• Connected component analysis
• Kalman-filter based tracking

Trajectory Analysis

• MSD calculations
• Velocity autocorrelation
• Directional persistence metrics

Statistical Classification

• Hypothesis testing
• Confidence interval estimation (❌ not implemented)
• Motility pattern classification