Image Enhancement in Spatial Domain Presented by Mr

Image Enhancement in Spatial Domain Presented by : Mr. Trushar Shah. ME/MC Department, U. V. Patel College of Engineering, Kherva

Today’s topics • • What is image enhancement? Approaches. Image processing in spatial domain. Implementation - Image negative - Contrast Stretching - Power law transformation - Dynamic range compression - Bit plane Slicing. - Gray level Slicing.

What is Image Enhancement? • To process an image so that the result is more suitable than the original image for a specific application. • Enhancement is the subjective process.

Approaches Image Enhancement Spatial Domain Point Processing Frequency Domain Filtering OR Masking

Approaches • Spatial domain – direct manipulation of pixel. • Frequency domain – Manipulation in frequency plane

Spatial domain x § Image can be modeled by a continuous function of two variables : (x, y) coordinates of point/pixel. § The image function values correspond to the brightness/intensity at image point and generally denoted by f(x, y).

Spatial domain(cont. ) • Point processing : Independent of neighbors • Masking : based on small sub image.

Image negative

N = Gmax - O

Contrast Stretching

Contrast Stretching • Factor that causes low contrast images § Lack of dynamic range. § Poor illumination • Algorithm • Implementation

Power law Transformations(g>1)

Power law Transformations(g<1)

Compression of dynamic range

Compression of dynamic Range • s = c. log(1+|r|) • Log function scales [0, 10^6] to [0, 6]. • c=255/6.

Bit plane slicing • Separating each bit from pixel gray level, and gathering same for all pixel will generate bit plane. • Monochrome images are made of the 8 -bit planes.

Gray level Slicing • Separating gray level range of interest to different level so that the region is highlighted.

Histogram • The histogram of a digital image with intensity levels in the range [0, L-1] is a discrete function h(rk)=nk where, - rk is the kth intensity value. - nk is the number of pixel with intensity rk. • Normalized Histogram: - A normalized histogram is given by p(rk) = nk/MN. - The sum of all components of normalized histogram is 1.

Histogram

Conclusion for Histogram Processing • The whole span of gray levels should be used. • Number of pixels for all the gray levels, should be equal. OR • The probability of occurrence of all gray level should be uniform.