We recognise shapes as identical despite rotation, scaling, or distortion

The Gestalt principle of invariance (shape constancy) describes the mind’s ability to identify an object as the same object across transformations: rotation, translation, size change, reflection, and even partial distortion. This is why we recognise faces from different angles, read letters in wildly different typefaces, and solve CAPTCHAs with stretched characters. For visual artists this is a key property of logomarks — a logo must be invariant across sizes, colours, and contexts. In generative art, breaking invariance by continuously deforming a shape creates perceptual tension because the brain cannot lock onto a stable identity.

Examples

CAPTCHA: distorted letters are still readable by humans because of invariance. A face recognised from an unusual angle. A wordmark stays readable at 10px and 100px.

Assessment

Take a simple geometric logo and transform it through five steps (rotate, scale, reflect, skew, fragment). At which step does recognition fail, and why?