What Enhancement Actually Does#

Enhancers modify the detect_mat accessor — the grayscale representation that detectors consume. They never touch rgb or gray, preserving the original image data for visualization and color measurements.

Why Enhance?#

Raw plate images often have properties that confuse detectors:

  • Noise — random pixel variation that creates false positives

  • Low contrast — faint colonies barely distinguishable from agar

  • Uneven illumination — brightness gradients across the plate

  • Texture — agar surface patterns that mimic colony edges

Enhancement addresses these problems by transforming detect_mat into a version where colonies stand out more clearly.

Categories of Enhancement#

Smoothing (Noise Reduction)#

Reduce random variation while preserving colony boundaries.

  • GaussianBlur — isotropic smoothing; fast but blurs edges

  • MedianFilter — removes salt-and-pepper noise; preserves edges better

  • BilateralDenoise — smooths within regions; preserves edges explicitly

  • StableDenoise (BM3D) — state-of-the-art block-matching denoising

Contrast Enhancement#

Increase the separation between colony and background intensities.

  • CLAHE — local adaptive histogram equalization; handles spatially varying contrast

  • ContrastStretching — linear remapping to fill the dynamic range

  • UnsharpMask — sharpens edges by subtracting a blurred version

Illumination Correction#

Remove large-scale brightness gradients.

  • HomomorphicFilter — frequency-domain separation of illumination and reflectance

  • SubtractGaussian — subtracts a heavily blurred background estimate

  • SubtractRollingBall — morphological background estimation

Structural Enhancement#

Enhance specific morphological features.

  • FrangiVesselness — enhances tubular structures (hyphae, branches)

  • SobelFilter — highlights edges

  • PhaseCongruencyEnhancer — illumination-invariant edge detection

Stacking Enhancers#

Enhancers compose linearly — each reads detect_mat, modifies it, and writes it back. Order matters:

  1. Denoise first — reduce noise before amplifying contrast

  2. Correct illumination — normalize brightness before thresholding

  3. Enhance contrast — maximize colony/background separation last

A typical preprocessing chain: GaussianBlur → HomomorphicFilter → CLAHE.

The Enhancement ↔ Detection Interface#

detect_mat  ──[Enhancer 1]──→  detect_mat  ──[Enhancer 2]──→  detect_mat
                                                                    │
                                                              [Detector]
                                                                    │
                                                              objmask, objmap

This clean interface means you can swap any enhancer without affecting the detector, and vice versa. The pipeline model makes this experimentation easy.