idiff

Idiff is a term used in image processing to refer to tools and algorithms that compare two images to identify and quantify differences. It is commonly employed in visual regression testing, quality assurance, and validation of image processing pipelines, where detecting even small changes between image versions is important.

Most idiff approaches involve an optional alignment or registration step to account for differences in position, scale, or rotation, followed by normalization of image size and color space. The core operation is a pixel-level difference that produces a diff image or a binary mask highlighting changed regions. Alongside the visual output, idiff tools typically report quantitative metrics such as mean absolute error, root mean square error, peak signal-to-noise ratio, or structural similarity index (SSIM) to summarize the magnitude of change.

Implementation details vary across platforms. Idiff can be provided as a standalone command-line utility, as a function within a library, or as part of larger image processing or testing frameworks. Consequently, the exact interface, supported image formats, and options for controlling sensitivity, thresholding, and output format differ between implementations.

Applications of idiff include automated web UI visual testing, comparison of successive image processing results, and quality control in media pipelines. By combining visual diffs with numerical metrics, these tools help users quickly assess whether changes are acceptable or require further investigation. See also image comparison, SSIM, PSNR, and pixel-wise difference.