• Visual image search engines exist only as experimental prototypes (demos, toys, etc). Worse yet, many make broad claims with nothing to back them up. If you have the technology, how hard is it to create a little web based demo!
• Most demos work with small collections of images, with no upload feature, which makes testing impossible.
• When testing is possible, the results are questionable. T
• he only well developed approaches are based on the distribution of colors, texture, and image segmentation.
• Fourier/wavelet transform has been mostly used for image compression.
• The engines provide only “likeness” search, which is completely subjective and obscures the fact the image analysis methods are inadequate.
• “User feedback”, “learning”, and “semantic” features obscure the fact the image analysis methods are inadequate.

Pixcavator 2.1. A couple of bugs have been fixed and a disabled slider has been enabled.

Photo album management. Based on the wavelet transform, which is simply a sophisticated averaging method. The best visual image search application I’ve found so far. Produces meaningful results. Based on solid mathematics. In fact, there is a paper describing everything. Of course, it’s still not exactly what I would like to see. Example  from the site: an image of sunset on a beach returns bright headlights in a city. Experiment: an image of a fingerprint is not matched with its rotated copy. The last one simply confirms what is discussed in the paper – the reliability of matching quickly deteriorates under rotations, distortions, etc.

Pixcavator 2 (389KB). There are still a couple of bugs and rough edges but this is a completely different level. Objects are listed in a table and squares are placed over them in the image. It is slower, so for the simplification tasks use version 1.9 below. To create watercolors from your photographs try Pixcavator Watercolor. SDK and documentation still to come…

Relies on color, texture, and image segmentation. Testing is currently unavailable. Example: an image of a leopard returns an image of a lion.