Photomosaics - Resolution Sample

As more and more images are used to make up the final image, the target image can be approximated better and better as it is roughly the equivalent of adding more and more pixels. In the following example, a picture of the head of Venus in Botticelli's famouse painting hanging at the Uffizi Gallery in Florence, Italy, is used as an example:

We begin by making a photomosaic out of 100 total tiles, in 10 rows and 10 columns. Obviously it is pretty hard to make out, but surprisingly enough if you step away from your computer screen and squint and know what the target image looks like you can already begin to see that details in the individual images are recognizable and have been picked very intelligently. Obviously a larger selection of images increases the odds that details in the target image can be matched. In this example a relatively limited set of 893 images was used.

The following photomosaic uses a total of 400 tiles, in 20 rows and 20 columns. A much larger amount of detail is visible in this example and it is clear that the image is starting to look like the target.

The following photomosaic uses a total of 1600 tiles, in 40 rows and 40 columns. The target image is being matched very well at this level.

The last photomosaic uses a total of 6400 tiles, in 80 rows and 80 columns. The target image is matched extremely well, and much smaller details are being captured. Look closely at the eyes, lips, nose and flowing hair for example.