Skip to content

README.md

Latest commit

 

History

History
89 lines (68 loc) · 3.91 KB

README.md

File metadata and controls

89 lines (68 loc) · 3.91 KB
Raw

Image API

Public API index | App guide

labkit.image.* provides GUI-free image file IO, conversion, preview sizing, and common enhancement operations. App-specific registration, measurement, ROI, matching, and export workflows build on these functions.

labkit.image.version() returns image-library version and compatibility information used by App definition.m requirement declarations.

Common Calls

filter = labkit.image.fileDialogFilter("IncludeAll", true);
paths = labkit.image.normalizePaths(rawPaths);
records = labkit.image.readFiles(paths);

source = labkit.image.im2double(records(1).image);
preview = labkit.image.ensureRgb(source);
preview = min(max(preview, 0), 1);
luma = labkit.image.rgb2gray(preview);
[preview, scale] = labkit.image.resizeToFit(preview, "MaxHeight", 1500);
[preview, budget] = labkit.image.previewBudget(preview, "MaxPixels", appPreviewBudget);

blurred = labkit.image.meanFilter2(preview(:, :, 1), 7);
enhanced = labkit.image.adjustBrightnessContrast(preview, 10, 20);
enhanced = labkit.image.localContrast(enhanced, 50, 3);
enhanced = labkit.image.sharpen(enhanced, 30, 1.5);
enhanced = labkit.image.grayWorldWhiteBalance(enhanced, 80, 0);

labkit.image.writeFile(enhanced, outputPath);

Normalization Helpers

labkit.image.im2double follows the MATLAB im2double call contract for supported numeric image classes, including the optional "indexed" mode.

labkit.image.rgb2gray follows the MATLAB rgb2gray call contract for RGB images and colormaps while using the documented Rec.601 luma transform.

labkit.image.ensureRgb changes channel shape only. It expands grayscale data to three channels or drops channels after RGB without changing class or sample values. Callers that need display-ready RGB data explicitly compose im2double, ensureRgb, and [0, 1] clamping as shown above.

labkit.image.previewBudget preserves native pixels by default. An App passes an explicit finite "MaxPixels" value only when its own workflow permits sampling; the resulting integer coordinate scale is returned with the preview.

Provided Operations

The module provides:

  • supported source-image extension lists and file-dialog filters
  • path normalization and display names
  • imread/imwrite wrappers that normalize app-facing edge behavior
  • MATLAB-compatible image conversion, explicit RGB shaping, preview-size fitting, and edge-normalized mean filtering
  • caller-owned display-pixel budget helpers for responsive previews while preserving a documented integer coordinate scale
  • generic image enhancement primitives such as brightness/contrast, HSV hue/saturation, gray-world white balance, local contrast, and sharpening

Applications add:

  • which filters or processing steps appear in the app UI
  • parameter labels, defaults, validation ranges, and step history semantics
  • reference matching, protected background/ROI behavior, scientific formulas, crop geometry, focus stacking, curvature, and DIC processing
  • export manifests, result columns, filenames, failed-row policy, alerts, and log text

Use labkit.image.readFiles when an app needs generic source-image records. Its option structure is a closed contract: unknown fields or a non-struct value raise labkit:image:InvalidOptions. Apps may copy the returned path, name, and image fields into their own item structures. Specialized formats and result structures remain documented by the app that uses them.

Reference Contract

The generated Image API pages linked from the public API index document exact syntax, inputs, outputs, implemented options and defaults, legal values, failure behavior, examples, and related functions. In particular, resizeToFit accepts only the documented "bilinear" and "nearest" methods and rejects any other method instead of silently selecting an interpolation policy.