CAPTURE AND TRANSPARENCY IN COARSE QUANTIZED IMAGES

This study examines the effect of coarse quantization (blocking) on im age recognition, and explores possible mechanisms, Thresholds for nois e corruption showed that coarse quantization reduces drastically the r ecognizability of both faces and letters, well beyond the levels expec ted by equivalent blurring, Phase-shifting the spurious high frequenci es introduced by the blocking (with an operation designed to leave bot h overall and local contrast unaffected, and feature localization) gre atly improved recognizability of both faces and letters, For large pha se shifts, the low spatial frequencies appear in transparency behind a grid structure of checks or lines, We also studied a more simple exam ple of blocking, the checkerboard, that can be considered as a coarse quantized diagonal sinusoidal plaid, When one component of the plaid w as contrast-inverted, it was seen in transparency against the checkerb oard, while the other remained ''captured'' within the block structure , If the higher harmonics are then phase-shifted by pi, the contrast-r eversed fundamental becomes captured and the other seen in transparenc y, Intermediate phase shifts of the higher harmonics cause intermediat e effects, which we measured by adjusting the relative contrast of the fundamentals until neither orientation dominated, The contrast match varied considerably,vith the phase of the higher harmonics, over a ran ge of about 1.5 log units, Simulations with the local energy model pre dicted qualitatively the results of the recognizability of both faces and letters, and quantitatively the apparent orientation of the modifi ed checkerboard pattern, More generally, the model predicts the condit ions under which an image will be ''captured'' by coarse quantization, or seen in transparency, (C) 1997 Elsevier Science Ltd.