Abstract

Most of the adaptive processes that have been proposed for image enhancement or data compression rely on some form of two-step mechanism characterized by an estimation of the local structure of the image, followed by an appropriately tuned filter to perform the desired operation. A variety of techniques have been used to perform the spatial adaptation. We review the major methods used for spatial activity indication to determine whether there is any particular method that provides superior performance. We present 14 methods of spatial adaptation and compare their performance in the presence of noise for both synthetic test images and real gray-level images. We conclude that the most effective indicator functions are those that include some implicit low-pass filtering to suppress random fluctuations and that are based on some form of gradient estimation.

© 1991 Optical Society of America

Dynamics of adaptation for vision with a stabilized image

B. E. A. Saleh, U. Tulunay-Keesey, J. N. Ver Hoeve, and M. Hom
J. Opt. Soc. Am. A 8(7) 1172-1181 (1991)

Filter for biomedical imaging and image processing

Partha P. Mondal, K. Rajan, and Imteyaz Ahmad
J. Opt. Soc. Am. A 23(7) 1678-1686 (2006)

Measurement of optical flow by a generalized gradient scheme

P. Sobey and M. V. Srinivasan
J. Opt. Soc. Am. A 8(9) 1488-1498 (1991)

Parallel image processing based on an evolution equation with anisotropic gain: integrated optoelectronic architectures

Mikhail A. Vorontsov
J. Opt. Soc. Am. A 16(7) 1623-1637 (1999)

Multiresolution, error-convergence halftone algorithm

Eli Peli
J. Opt. Soc. Am. A 8(4) 625-636 (1991)