Encoding Digital Images

Grade Level: 8th - 12th; Type: Computer Science

Objective:

The goal of these experiments is to learn about different types of digital images, compare different encoding algorithms used to compress the data, and carefully evaluate the quality of lossy vs. lossless compression. Students will print out images that are compressed using different encoding algorithms, and evaluate which works best.

Research Questions:

• Do JPEG, TIFF, PNG and GFF rely on lossy or lossless compression?
• What is Huffman coding?
• How is sound compressed?
• Which types of compression algorithmns work well for color images? Very detailed images? Images of the out-of-doors?

Whether you are taking photos with your iPhone, watching a DVD, reading a pdf, or checking out Google images, digital images are everywhere. This is quite amazing because these images require a lot of memory. The memory required by these complex images makes them hard to store and time-consuming to download.

Almost everyone is familiar with pixels – the rows of tiny dots that make up an image. The resolution of an image depends upon the number of pixels used in a particular image. A raster image is a data structure that stores information about those pixels. The data about these pixels is compressed, using a particular encoding algorithm so that the image takes up less memory. A lossless algorithm permits the image to be reproduced without losing any information. A lossy algorithm is one in which some of the data is lost so that the image can be compressed even more.

Materials:

• Computer
• Printer
• Assistant
• Internet access or a library of different photo images stored on the computer.

Experimental Procedure:

1. Select at least two images. The first image should be a nature image with expansive blue sky, ocean or other repetitive information. The second image should be a busy urban scene. If you select a third image, it should be black and white.
2. Determine how to save your images in difference formats. For example, if you are working in MAC OS X, find a image in your document. Open the document in “Preview.” Click on “File” to open the drop-down menu. Click on “Save As.” Select a particular format (TIFF, pdf, JIF, PNG etc.) under “Format” and then click on “Save.”
3. Save the same image in a variety of formats.
4. Have your assistant print the image. The assistant should mark the image with a code number that is associated with the format. For example, all jpegs can be marked “#1,” all TIFFs can be marked “2,” and so on. Only the assistant knows what formats are associated with which numbers.
5. Closely examine the printed images. Pay close attention to the fine details of the image. Are all the images identical? Which are better?
6. Research which formats involve lossy coding and which involve lossless coding.

Terms/Concepts: Lossy vs. lossless; Data compression; Encoding; Resolution; Pixel; Raster; Bitmap; jpg, pdf. TIFF GIF; Huffman coding

References:

• Raster Graphics http://en.wikipedia.org/wiki/Raster_graphics
• Huffman Coding http://en.wikipedia.org/wiki/Huffman_coding
• Trees for Data Compression http://www.programmersheaven.com/2/Art_Huffman_p1
• Lossless vs. Lossy Compression http://www.maximumcompression.com/lossless_vs_lossy.php