batch_crop.cropGeometry.cropImage

Description

Adds optional padding, rotates the resulting canvas, and extracts a crop centered at a point expressed in the original image coordinate system. The crop always has the requested pixel dimensions. Areas outside the transformed canvas use the selected fill value. This function does not resize the crop, display graphics, or write files.

Inputs

imageData

Nonempty numeric or logical image. Accepted sizes are M-by-N and M-by-N-by-C. The output preserves its class and channel count.

opts

Scalar structure containing the crop options below.

Options

cropWidth

Required positive pixel count. Noninteger values are rounded.

cropHeight

Required positive pixel count. Noninteger values are rounded.

angleDeg

Counterclockwise canvas rotation in degrees. Default: 0.

centerXY

Crop center [x y] in original-image pixel coordinates, where x is the column and y is the row. Invalid or empty input selects the geometric image center. The center is limited so the requested crop remains representable on the prepared canvas. Default: [].

paddingPercent

Padding added around each source edge as a percentage of the corresponding source dimension. Negative values are treated as zero by the canvas preparation code. Default: 0.

fillValue

Scalar or channel-compatible value used outside the rotated image. Empty input selects the app's white value for imageData. Default: [].

Outputs

result

Scalar crop result structure. result.image is exactly cropHeight-by-cropWidth-by-C. result.centerX and centerY retain the effective center in original-image coordinates. rotationDeg, paddingPercent, sourceWidth, sourceHeight, cropWidth, cropHeight, nativeCropWidth, nativeCropHeight, outputWidth, and outputHeight describe the operation. scaleMode is "Pixels", ok is true, status is "cropped", and message is "OK".

Errors

labkit_BatchImageCrop_app:InvalidImage

imageData is empty, nonnumeric, nonlogical, or has more than three dimensions.

labkit_BatchImageCrop_app:InvalidCropSize

cropWidth or cropHeight is absent, nonfinite, nonscalar, or smaller than one.

Example

imageData = reshape(uint8(1:100), 10, 10);
opts = struct("cropWidth", 4, "cropHeight", 3, "centerXY", [5 6]);
result = batch_crop.cropGeometry.cropImage(imageData, opts);
assert(result.ok && isequal(size(result.image), [3 4]))