1. CS 128/ES 228 - Lecture 7a1 Data – How (Much of) It Is Stored

2. CS 128/ES 228 - Lecture 7a2 Outline What Is an Image Really? Methods of Storing Images How to Make a Big File Small Compression Algorithms Conversion Algorithms In theory In practice

3. CS 128/ES 228 - Lecture 7a3 What is an image? An image is anything we store on the computer that we think of as a “picture”. It should look “the same” on any display. Image file formats GIF, JPEG, TIFF, BMP NOT shapefiles

4. CS 128/ES 228 - Lecture 7a4 File Formats There are many image file formats 35 on the first page I hit looking for a list! Each has advantages and disadvantages

5. CS 128/ES 228 - Lecture 7a5 GIF Developed by Compuserve in 1987 Particularly good for line drawings (anything with sharp edges) VERY common on web

6. CS 128/ES 228 - Lecture 7a6 JPEG (or JPG) Product of the Joint Photographers Experimental Group Good for photos, images with subtle changes Also popular on the web

7. CS 128/ES 228 - Lecture 7a7 GIF vs. JPEG JPEGGIF Use For“Realistic” artworkIllustrations Compres- sion Lossy, but controllableLossy, no control Colors24-bits8-bits OthersNo transparencyTransparency

8. CS 128/ES 228 - Lecture 7a8 JPEG 2000 (aka JP2) “The JP2 and JPX file formats allow for handling of color-space information, metadata, and for interactivity in networked applications as developed in the JPEG Part 9 JPIP protocol.” Some imagery is now distributed as JP2 files – datum and projection included at no extra charge!

9. CS 128/ES 228 - Lecture 7a9 Portable Network Graphics (PNG) PNG also stands for “PNG’s Not GIF” Loss-less compression using non- patented algorithm Supports transparency, but not really animation ISO standard since 2003

10. CS 128/ES 228 - Lecture 7a10 BMP Bitmap format – Primarily for Windows (but not exclusively) NO Compression means LARGE files Standard Screen Snapshot is BMP

11. CS 128/ES 228 - Lecture 7a11 EPS, PICT, TIFF Encapsulated PostScript (mostly for printing, some display) PICTure format (Macs only) Tag Interchange File Format (multi- platform, but less used these days)

12. CS 128/ES 228 - Lecture 7a12 Shapefiles and active software A running program may read from or write to these formats, but generally uses its own memory management while running. Shapefiles contain shape information and are not in any of these formats – and not truly image files They are vector layers, after all

13. CS 128/ES 228 - Lecture 7a13 Compression Algorithms Compression algorithms “shrink” files May do so by mathematical “tricks” or by discarding information

14. CS 128/ES 228 - Lecture 7a14 Two KEY Facts about Compression NO LOSS-LESS compression algorithm can work all the time! NO LOSSY compression algorithm can regenerate its original data.

15. CS 128/ES 228 - Lecture 7a15 An LOSS-LESS Example Run-length compression Count and record the length of the data set and then each group of 0’s or 1’s 1110100 1110000 1000000 3 7 0 3 1 1 2 3 4 1 6

16. CS 128/ES 228 - Lecture 7a16 A LOSSY Example Truncation 1242144903 0293570214 9352109521 7259027565 3048282535 124 029 935 725 304 1240000000 0290000000 9350000000 7250000000 3040000000

17. CS 128/ES 228 - Lecture 7a17 How much does compression affect image quality? Original (32 MB) Compressed (493 kB)

18. CS 128/ES 228 - Lecture 7a18 Converting Vector to Raster Must compute the equation of the line Then choose which pixels to highlight Many algorithms, but differences are technical

19. CS 128/ES 228 - Lecture 7a19 Typical algorithm Y = y0 + 1 Illuminate pixel (x, int(Y)) Y = Y + 1 X = X + 1 /m Illuminate pixel (x, int(Y)) … Until Y == y1 X = x0 Y = y0 Illuminate pixel (x, int(Y)) (x1,y1) (x0,y0)

20. CS 128/ES 228 - Lecture 7a20 Anti-aliasing Basic idea – Remove the “jaggies” by using color variations

21. CS 128/ES 228 - Lecture 7a21 Conversion in practice

22. CS 128/ES 228 - Lecture 7a22 Converting Raster to Vector Basic idea Find areas with sharp changes – these are your boundaries. Adjust as topology indicates Much harder in practice than the other way around Alternative is hand-digitization