Mahdi Amiri. March Sharif University of Technology

Transcription

1 Course Presentation Multimedia Systems Image II (Image Enhancement) Mahdi Amiri March 2014 Sharif University of Technology

2 Image Enhancement Definition Image enhancement deals with the improvement of visual appearance of the scene, to improve the detectability of objects to be used by either a machine vision system or a human observer. Sources of image deterioration Noise Low Resolution Quantization levels Source of noise Note: Enhancement vs. Restoration Image enhancement is the improvement of digital image quality (wanted e.g. for visual inspection or for machine analysis), without knowledge about the source of degradation. If the source of degradation is known, one calls the process image restoration. Here, we are not going to specifically differentiate these techniques. Electronic signal fluctuations in detector ( CCD chip ) Page 1

3 Image Enhancement Have seen so far Gamma Correction Histogram Equalization Page 2

4 Image Enhancement Image Noise The random variation of brightness or color information in images An undesirable by-product of image capture grainy image, noise from a digital camera Image with salt and pepper noise Page 3

5 Image Filters Gaussian smoothing Noise fluctuations are rapid, ie, high frequency. Gaussian filters are a class of smoothing filters where the kernel values have a 2D Gaussian shape. 1D Gaussian Filter 1/16 Page x3 Kernel 2D Gaussian Filter 1/256 Sweep The Image x5 Kernel

6 Image Filters Mean Filter To replace each pixel value in an image with the mean (`average') value of its neighbors Kernel: represents the shape and size of the neighborhood 1/9 1/9 1/9 1/9 1/9 1/9 1/9 1/9 1/9 Typ. Mean Filter Kernel A halftone print Mean filtered Page 5

7 Image Filters Median Filter Replacing each entry with the median of neighboring entries Nonlinear digital filtering technique Used to remove noise Pro: under certain conditions, it preserves edges while removing noise 1D Example: Input: x = [ ], Window size: 3 The median filtered output signal y: y[1] = Median[2 2 80] = 2 (Left padding with 2) y[2] = Median[2 80 6] = Median[2 6 80] = 6 y[3] = Median[80 6 3] = Median[3 6 80] = 6 y[4] = Median[6 3 3] = Median[3 3 6] = 3 (Right padding with 3) i.e. y = [ ]. Page 6

8 Median Filter 2D Median Filter Example Page 7 Input Image Mean Filtered Image Median Filterd Image

9 Box Filtering Original unfiltered image Kernel Page 8

10 Box Filtering public int[] BoxFiltering( int[] pixels,int width, int height, float[] kernel) { int[] temp = new int[width*height] ; float denominator = 0.0f ; float red, green, blue ; int ired, igreen, iblue, indexoffset, rgb ; int[] indices = { -(width + 1), -width, -(width - 1), -1, 0, +1, width 1, width, width + 1 } ; for (int i=0;i<kernel.length;i++) denominator += kernel[i] ; if (denominator==0.0f) denominator = 1.0f ; for (int i=1;i<height-1;i++) { for (int j=1;j<width-1;j++) { [Include Part A] } } return temp ; } See interactive-applet/ applet/ Example Java Code [Part A] red = green = blue = 0.0f ; indexoffset = (i*width)+j ; for (int k=0;k<kernel.length;k++) { rgb = pixels[indexoffset+indices[k]] ; red += ((rgb & 0xff0000)>>16)*kernel[k] ; green += ((rgb & 0xff00)>>8)*kernel[k] ; blue += (rgb & 0xff)*kernel[k] ; } ired = (int)(red / denominator) ; igreen = (int)(green / denominator) ; iblue = (int)(blue / denominator) ; if (ired>0xff) ired = 0xff ; else if (ired<0) ired = 0 ; if (igreen>0xff) igreen = 0xff ; else if (igreen<0) igreen = 0 ; if (iblue>0xff) iblue = 0xff ; else if (iblue<0) iblue = 0 ; temp[indexoffset] = 0xff ((ired<<16) & 0xff0000) ((igreen<<8) & 0xff00) (iblue & 0xff) ; The parameter pixels is an array containing a total of width*height pixel information. The parameter kernel is an array with a fixed length of nine. This is because the implementation assumes a window of size 3x3. The function also assume each pixel is in the form ARGB (alpha, red, green, blue), where blue is the least significant byte. Alpha is the transparency information and should just be left untouched. The array indices is a simple optimization table used to find neighboring pixels. The denominator is the sum of parameter kernel, it will be the denominator when calculating average. Page 9

11 Box Filtering Smoothing (~ Mean Filter) Kernel (with denominator) 1/10 1/10 1/10 1/10 2/10 1/10 1/10 1/10 1/10 Kernel (without denominator) Page 10

12 Box Filtering Sharpening (Edge Enhancement) Kernel (without denominator) Page 11

13 Box Filtering Raised Kernel (without denominator) Page 12

14 Box Filtering Motion Blur Kernel (without denominator) Page 13

15 Box Filtering Edge Detection Kernel (without denominator) Page 14

16 Image Enhancement Despeckle Speckle detection and deletion. Allows the removal of speckle in scanned or faxed images. The speckle is the presence of black points of noise in images acquired by a scanner or received by fax. Page 15

17 Image Enhancement Ref.: ikpe1101.ikp.kfa-juelich.de/briefbook_data_analysis/node178.html Morphological Operations Mathematical morphology is a set-theoretical approach to multidimensional digital signal or image analysis, based on shape. The signals are locally compared with so-called structuring elements S of arbitrary shape with a reference point R. We will define these operations without any mathematical rigour, For a binary image: Erosion: The eroded image of an object O with respect to a structuring element S with a reference point R, O S, is the set of all reference points for which S is completely contained in O. Dilation: The dilated image of an object O with respect to a structuring element S with a reference point R, O S, is the set of all reference points for which O and S have at least one common point. Opening: Is defined as an erosion, followed by a dilation: (O S) S Closing: Is defined as a dilation, followed by an erosion: (O S) S structuring element Sample structuring elements Page 16

18 Morphological Operations A Simple Example Structural element (SE) Erosion with Z 8 Dilation with Z 8 Other example application: Major preprocessing step for object tracking or optical character recognition. Page 17

19 Morphological Operations Example Application Color filtering based object tracking application example original frame (zoomed-out) Page 18 HSV thresholded image H: , S: , V: After erosion and dilation (opening) Erosion SE: 3 3 Rect. (Z 8 ) Dilation SE: 8 8 Rect. (Z 8 ) Now the desired object can be detected by finding the largest closed area.

20 Color Adjustments Example of color correction in Photoshop Page 19 White balancing method for color correction is shown in the next slid.

21 Note: 25 C (Celsius ) (about 77 Fahrenheit or 298 Kelvin) White Balancing Color Temperature Incorrect white balance Correct white balance Color Temperature comparison of common electric lamps (Warmer light ~ Lower color temperature) Color Temperature: Based on the color given off by a glowing hot piece of platinum. When moving from a bright daylight environment to a room lit by a candle, our brain can quickly adjust to the changes, making white appear white, whereas film is balanced for one particular color and anything that deviates from this will produce a color cast. Example for the psychological effects of color: A warmer (i.e., lower color temperature) light is often used in public areas to promote relaxation, while a cooler (higher color temperature) light is used to enhance concentration in offices. Page 20 Using portable references for white balancing

22 Color Temperature The Psychology of Color While perceptions of color are somewhat subjective, there are some color effects that have universal meaning. Office: Blue, Most productive color. Bedroom: Green, Tranquility and Health. Girl's Room: Pink, Calming, Warm. Kitchen: Yellow, Increases metabolism, brightens room, gives you energy. Living Room: Lavender (pale purple), Calms the nerves, allows relaxation. Dining Room: Red, Encourages appetite. Page 21 Source: /how-to-choose-a-color-scheme-for-your-wordpress-theme/

23 The Psychology of Color Page 22

24 Image Enhancement Application Real-time Filtering HD video conferencing, multi-focus imaging Page 23

25 Image Editors Paint.NET Paint.NET is a proprietary freeware raster graphics editor program for Microsoft Windows, developed on the.net Framework. Page 24

26 Image Editors GIMP GIMP (short for the GNU Image Manipulation Program) is a free software raster graphics editor. (Microsoft's Windows, Apple's Mac OS X, GNU/Linux) GIMP running under X11 on Mac OS X GIMP 2.6 running on Ubuntu See: Page 25

27 Multimedia Systems Image II Thank You Next Session: Image III FIND OUT MORE AT Page 26