SIFT

Sift is a word with several related senses. As a verb, it means to separate and remove larger particles from a mixture by passing it through a sieve or mesh. It can also mean to examine something carefully in order to sort or select items, for example, to sift through records or evidence. As a noun, a sift is a device—typically a small frame with a mesh screen—used to perform this operation.

In cooking and food preparation, sifting is used to aerate dry ingredients such as flour, cocoa powder, or powdered sugar and to break up clumps. Kitchen sifters include hand-held sifters, sifter baskets, and attachments for mixers. Sifting improves texture and measurement accuracy in some recipes.

In computer vision, SIFT stands for Scale-Invariant Feature Transform. Introduced by David Lowe in 1999, it describes a method to detect and describe local features in images that are stable under changes of scale, rotation, and illumination. SIFT features enable tasks such as image matching, object recognition, panorama stitching, and 3D reconstruction. The algorithm has been widely cited and has influenced many successor methods; patents historically restricted its use in some contexts, leading to alternative descriptors in open implementations.

The term sift is also used more broadly to indicate filtering or sorting large volumes of information or materials, for example, to sift through data or sift a supply chain for quality. The usage ranges from everyday kitchen tasks to formal analyses and technical literature.

See also: sieve; flour sifting; scale-invariant feature transform.