Media Processing

The Media Processing Widget provides a collection of tools to adjust, enhance, and transform media data (images or video) within the viewer.
Each processing option can be enabled by checking the Active box, and its effect will be applied directly in the media viewer.

This guide introduces each option, its background, and how to use each effectively.

When media is selected in the feature table within the media widget, you will be presented with the following actions:

Linear Transform

A linear transform modifies pixel intensity values by applying a simple linear function:

\[ I_{out}(x, y) = \alpha \cdot I_{in}(x, y) + \beta \]

  • Alpha (Contrast): Controls contrast. Values >1 increase contrast, values between 0 and 1 decrease contrast.
  • Beta (Brightness): Adjusts brightness by shifting all pixel values up or down.
  • Minimum / Maximum: Sets intensity clipping or scaling boundaries. Pixels outside this range are clamped. This is often used for rescaling pixel intensities into a normalized display range.

Use case: Brightening dark images, increasing contrast, or standardizing pixel intensity ranges.

Gamma Correction

Gamma correction is a nonlinear intensity adjustment:

\[ I_{out}(x, y) = I_{in}(x, y)^\gamma \]

  • Gamma: The power-law exponent.
    • γ < 1: Brightens images (boosts darker regions).
    • γ > 1: Darkens images (suppresses bright regions).

This compensates for human visual perception, which is more sensitive to relative changes in darker tones.

Use case: Making hidden details in dark or bright regions more visible.

Image Sharpening

Sharpening highlights edges by enhancing high-frequency details. This method uses an unsharp mask:

\[ I_{out} = I_{in} + \lambda \cdot (I_{in} - G_\sigma(I_{in})) \]

Where \(G_\sigma\) is a Gaussian blur applied with standard deviation σ.

  • Sigma: Controls the blur radius used before sharpening. Small values sharpen fine details; larger values enhance broader structures.

Use case: Highlighting boundaries, edges, and fine structures in images.

CLAHE (Contrast-Limited Adaptive Histogram Equalization)

CLAHE improves local contrast by redistributing pixel intensities based on local histograms. Unlike global histogram equalization, CLAHE works in small tiles (grid regions).

  • Clip Limit: Prevents noise amplification by limiting the contrast enhancement.
  • Grid Size: Defines the size of the local regions used for histogram equalization.

Use case: Bringing out local details in images with varying illumination, e.g., microscopy or medical images.

Bilateral Filter

The bilateral filter smooths images while preserving edges. It considers both spatial proximity and intensity similarity:

\[ I_{out}(x, y) = \frac{1}{W} \sum_{i,j} G_s(||p-q||, d) \cdot G_r(|I(p)-I(q)|, \sigma_r) \cdot I(q) \]

  • Diameter: Size of the filter kernel (neighborhood).
  • Color Sigma (σr): How much intensity differences are preserved; large values allow smoothing across intensity changes, small values preserve edges.
  • Space Sigma (σs): How far pixels influence each other in space.

Use case: Noise reduction while keeping edges sharp (useful for video or structural microscopy data).

Median Filter

The median filter replaces each pixel with the median value of its neighborhood.

  • Kernel Size: Defines the neighborhood size. Larger kernels smooth more but may lose detail.

Use case: Removing salt-and-pepper noise or impulse noise.

Magic Eraser

The magic eraser provides an interactive editing tool for masking unwanted regions.

  • Brush Size: Size of the drawing tool when marking regions.
  • Median Filter Size: Strength of smoothing applied to the mask.
  • Start Drawing Button: Enables drawing mode inside the media viewer.
  • Clear Mask Button: Resets the mask.

Use case: Manually erasing artifacts, covering distracting elements, or excluding regions from processing.

Colormap

A colormap maps grayscale intensity values to colors using a lookup table (LUT). This improves visibility and interpretation of data.

  • Lookup Table (LUT): Defines the mapping (e.g., Jet, Hot, Cool). Each intensity is assigned a unique color.
  • Alpha (Blend): Controls transparency of the colormap overlay (0 = fully transparent, 1 = fully opaque).
  • Normalize Checkbox: Rescales intensity values into a normalized range before applying the colormap, ensuring full use of the LUT range.

Use case: Enhancing interpretability of microscopy, behavior, or heatmap-like data.

Summary Table

Processing Option Key Parameters Purpose
Linear Transform Alpha, Beta, Min, Max Contrast & brightness adjustment
Gamma Correction Gamma Nonlinear intensity adjustment
Sharpening Sigma Enhance edges and details
CLAHE Clip Limit, Grid Size Local contrast enhancement
Bilateral Filter Diameter, Color Sigma, Space Sigma Edge-preserving smoothing
Median Filter Kernel Size Noise removal
Magic Eraser Brush Size, Median Filter Size Manual region masking
Colormap LUT, Alpha, Normalize Visual interpretation