Enlarges an Image by an integer factor using a classical CPU interpolation kernel (nearest, bilinear, bicubic, or Lanczos) and emits the resized Image; fast and detail-preserving but not generative.
Typical backends
Cheaply enlarge a low-resolution feed and publish the result: a file source feeds frames, this component resizes them with a classical kernel, and the output is streamed over HTTP.
Upscale small text crops before OCR so the recognizer sees larger glyphs, then publish the recognized text as JSON.
Caveats
- LimitationClassical interpolation resamples existing pixels and cannot invent detail; pair with a neural super-resolver when reconstructing genuinely lost high-frequency content matters.
- Parameter interactionThe minimum-width and minimum-height floors gate the skip TOGETHER: upscaling is skipped only when every non-zero floor is already satisfied, so an image already at or above the target size is passed through untouched. A floor of zero removes the constraint on that dimension.
- I/O contractThe factor is an integer multiplier; the output can be clamped to maximum dimensions after enlarging, and the result is a same-color-space
Image. - CostInference runs entirely on CPU with no model download, so it is cheap and has no GPU requirement, making it a good default when generative quality is not needed.