Multimedia Systems Image I (Acquisition and Representation) Mahdi Amiri November 2015 Sharif University of Technology

Transcription

1 Course Presentation Multimedia Systems Image I (Acquisition and Representation) Mahdi Amiri November 2015 Sharif University of Technology

2 Quantization Levels Image Representation Color Depth The number of bits used to represent the color of a single pixel. bits per pixel (bpp). 1bit: Monochrome 24bit: Truecolor 1bit 2bit 4bit 8bit 24bit Page 1

3 Image Representation Indexed Color, Palette It is a form of vector quantization compression. A 2-bit indexed color image Color table (the palette) 8-bit (256-color) Indexed image and its own palette 8-bit Grayscale image and palette Page 2

4 Image Representation, Palette Disadvantages Limited set of simultaneous colors per image. If the original color palette for a given indexed image is lost, it can be nearly impossible to restore it. 24-bit 8-bit 4-bit Incorrect palette Page 3

5 Image Representation Halftone A technique that simulates continuous tone imagery through the use of dots, varying either in size, in shapeor in spacing. Color Halftoning Page 4 Halftone dots How the human eye would see this sort of arrangement from a sufficient distance. Three examples of color halftoningwith CMYK separations. From left to right: The cyan separation, the magenta separation, the yellow separation, the black separation, the combined halftone pattern and finally how the human eye would observe the combined halftone pattern from a sufficient distance.

6 Image Representation, Dithering Definition An intentionally applied form of noiseused to randomize quantization error. Etymology: Mechanical computers performed more accurately when flying on board the aircraft, and less well on ground! Application: Increasing color depth without adding new bits 24-bit 1-bit black and white thresholding 1-bit, with Floyd-Steinberg dithering Page 5

7 Image Representation, Dithering Floyd Steinberg Algorithm Distribute the quantization residual to neighboring pixels that have not yet been processed. Pseudocode: for each y from top to bottom for each x from left to right Distribution matrix oldpixel := pixel[x][y] newpixel := find_closest_palette_color(oldpixel) pixel[x][y] := newpixel quant_error := oldpixel newpixel pixel[x+1][y] := pixel[x+1][y] + 7/16 * quant_error pixel[x-1][y+1] := pixel[x-1][y+1] + 3/16 * quant_error pixel[x][y+1] := pixel[x][y+1] + 5/16 * quant_error pixel[x+1][y+1] := pixel[x+1][y+1] + 1/16 * quant_error Page 6

8 Image Representation, Dithering Color Banding Artifact Dithering prevents large-scale patterns such as "banding" in images. Web-safe color palette with no dithering Web-safe color palette with Floyd Steinberg dithering Page 7

9 Image Resolution, Most Common Display Resolutions Image Resolution HD Image Resolution 3 rd gen. of ipad (QXGA) Google's Nexus 10 (WQXGA) Full-HD Ref.: en.wikipedia.org/wiki/display_resolution Page 8

10 Image Resolution The term 4K refers to the horizontal resolution of these formats, which are all on the order of 4,000 pixels. 4KUltra high definition television (UHD) is a resolution of 3840pixels 2160pixels (8.3 megapixels, aspect ratio 16:9). 8KUHD which is 7680pixels 4320 pixels (33.2 megapixels). Digital TV Resolutions 16:9 resolutions in comparison. en.wikipedia.org/wiki/4k_resolution Page 9

11 Image Resolution UHDTV, HDTD, SDTV SDTV: Standard-definition television. HDTV: high-definition television. UHDTV: Ultra HDTV. Chart showing resolutions for 8K UHDTV, 4K UHDTV, 1080p HDTV, and 480i SDTV. Diagram of the CIE 1931 color space that shows the Rec (UHDTV) color space in the outer triangle and Rec. 709 (HDTV) color space in the inner triangle. Both Rec and Rec. 709 use Illuminant D65 for the white point. Page 10

12 Image Resolution One million pixels To express: Megapixel (MP) The number of pixels in an image The number of image sensor elements of digital cameras The number of display elements of digital displays sensor elements, or QXGA display 3.1 MP ( = 3,145,728) 8 MP Phone Camera 160 MP Camera Page 11

13 Image Resolution Pixels per inch (ppi) Pixels per inch (PPI) or pixel density is a measurement of the resolution of devices in various contexts; typically computer displays, image scanners, and digital camera image sensors. 18 ppi 72 ppi 150 ppi Page 12

14 Image Resolution Pixels per inch (ppi) The average human eye can only detect 300 ppi. iphone 6s 4.7" 1334x ppi iphone 6s Plus 5.5" 1920x ppi ipad Pro 12.9" 2732x ppi Lumia " 2560x ppi Galaxy Note 5 5.7" 2560x ppi LG G4 5.5" 2560x ppi List of displays by pixel density Page 13

15 List of displays by pixel density Image Resolution, ppi Pixels per inch for a few more devices. iphone 5/5s 4" 1136x ppi ipad 4/Air 9.7" 2048x ppi HTC One 4.7" 1920x ppi Lumia " 1280x ppi Galaxy Note 3 5.7" 1920x ppi Xperia Z 5" 1920x ppi iphone 4, 4s 3.5" 960x ppi ipad 1, ipad 2 9.7" 1024x ppi ipad 3 9.7" 2048x ppi Nokia N95 2.6" 320x ppi Nokia Lumia " 800x ppi Galaxy Note II 5.55" 1280x ppi Google Nexus One 3.7" 800x ppi Samsung I9100 Galaxy S II 4.27" 800x ppi Page 14

16 Image Vision Histogram Plots the number of pixels for each tonal value. By looking at the histogram for a specific image a viewer will be able to judge the entire tonal distribution at a glance. Count (Number of pixels for each different intensity value) Image histogram Intensity (tonal value) Page 15

17 Image Vision Formula Contrast Contrast is the difference in visual properties that makes an object distinguishable from other objects and the background. Typ. histogram of a low contrast image Typ. histogram of a high contrast image Page 16

18 Image Vision Histogram equalization is a method in image processing of contrast adjustment using the image's histogram. Histogram Equalization Page 17

19 Image Vision Histogram equalization, Example Page 18

20 Image File Formats Raster and Vector Graphics Raster Graphics (Bitmap).BMP,.JPG,.PNG,.GIF Both.AI,.CDR,.PSD,.TIFF Vector Graphics.CGM,.SVG Page 19

21 Image Representation Panorama Example Software: Hugin and AutoStitch Stitching images captured above Milad Tower Page 20

22 Image Representation AutoStitch Process Example algorithm: SIFT Keypoint detection and matching. Page 21

23 Image Representation Large Screen Projection Page 22

24 Image Acquisition High-Dynamic-Range (HDR) 4 Images captured with different Exposure Values (EV or stop) HDR, Accurately representing the range of intensity levels found in real scenes Page 23

25 Image Acquisition HDR Movie Demo Play HDR movie demo Page 24

26 Image Acquisition, HDR Algorithm: Tone Mapping This is Exposure Bracketing Tone mapped HDR image To overcome the limited dynamic range of current standard digital imaging techniques A simple version of tone mapping: Mean Value Mapping Page 25

27 Image Acquisition Focus Bracketing Bracketing is the general technique of taking several shots of the same subject using different camera settings. A sequence of 5 incrementally focused images Example Software: CombineZP Page 26 Focus stacked image

28 Image Acquisition Focus Bracketing Example Application: Microscopy The three source image slices at three focal depths Contributions in the final "focus stacked" image The resulting focus stacked image with an extended depth of field Page 27

29 Multimedia Systems Image I Thank You Next Session: Image II FIND OUT MORE AT Page 28