For this activity, we want to try our hands on image compression. There are two main ways to compress an image - lossy and lossless compression - and there are many techniques underneath both. There is discrete cosine transform [
1], color pallete indexing [
2] and chroma subsampling [
3] for lossy compression. On the other hand, there is run-length encoding [
4] and DPCM [
5] for lossless compression. For more details on these types of compression, you can read the links provided.
Here, we used principal component analysis (PCA) to compress an image. First, I chose a simple image, one from the Windows 7 sample images. I resized the original 1024x768 into a 800x600 image so that processing will be a little faster.
Figure 1. "Tulips" by David Nadalin, 2008. From Windows 7
default sample images.
