Zero point operations in image and video processing? Sem 6 AiDs

Zero-memory operations (point operations) are image enhancement operations in which each output pixel depends only on the grey level of the same input pixel, \(s = T(r)\), not on its neighbours or on other frames; examples are negative, thresholding, contrast stretching, clipping, grey-level slicing, bit-plane slicing, log and power-law transformations.

They are called "zero memory" because the system does not need to remember any other pixel values: a 1×1 neighbourhood is enough. They can be implemented with a look-up table of \(L\) entries (\(L = 256\) for 8-bit images), which makes them very fast.

Common zero-memory operations (\(r\) = input grey level, \(s\) = output, range 0 to \(L-1\)):

1. Identity: \(s = r\).
2. Image negative: \(s = (L-1) - r\). For 8-bit images, \(r = 50\) becomes \(s = 205\). Useful for viewing white details in dark regions, e.g. mammograms.
3. Thresholding: \(s = L-1\) if \(r \ge T\), else 0. Converts a grey image to binary (e.g. text from background).
4. Contrast stretching: a piecewise-linear function with break points \((r_1, s_1)\) and \((r_2, s_2)\) that expands the range of grey levels in the region of interest, improving low-contrast images.
5. Clipping: a special case of contrast stretching where grey levels below \(a\) and above \(b\) are set to fixed values (slopes outside \([a, b]\) are zero).
6. Grey-level (intensity-level) slicing: highlights a range \([A, B]\), either making it bright and everything else dark (without background), or making it bright and keeping other levels unchanged (with background).
7. Bit-plane slicing: separates the 8 bits of each pixel into 8 binary images; higher-order planes hold most of the visual information, useful for analysis and compression.
8. Log transformation: \(s = c\log(1 + r)\). Expands dark values and compresses bright ones; used to display Fourier spectra.
9. Power-law (gamma) transformation: \(s = c\,r^{\gamma}\). \(\gamma < 1\) brightens dark images, \(\gamma > 1\) darkens; used for gamma correction of displays.

Contrast with other operations: neighbourhood (spatial) operations such as mean or median filtering use the pixels around a point (they have "memory" of neighbours); temporal operations use pixels from previous frames in video.