There is a real appeal for watching pictures that have been transformed somehow by an algorithm. In particular, these two pictures I took in the French Alps which were connected through the ASIFT algorithm.
Cable and I were discussing this and we figured that more examples were a way for certain techniques to be embraced or at the very least be part of others' mental process. It may be obvious in certain circles like image processing or even machine learning, but the integration of new algorithms, especially when it comes to video, has certainly not crossed over to real engineering and even some aspect of physics. More generally, the current thinking by non science practitioners is that your local police department can do CSI duties on a whim with the latest and sometimes unreal image enhancement capabilities or that most of the CGI in the latest summer blockbuster must have some truth. We are going to start a series of entries called CAI for Cable And Igor's adventures that will feature real examples of what some of the most advanced signal processing or matrix factorization techniques can do. We recently started with some of the videos taken by an endoscopic camera in reactor 5 of Fukushima Daichi.
The next entry will have a new situation. Part of this process is for us to make sense of what we see. In the videos above, we really wanted to see a convolution ( It never was noise; Just a different convolution )
If you have some time and interested in this experiment of using YouTube videos, we'll link to your transformed videos. Mesmerize us. In the meantime, we'll go through some selection of video excerpts and how they were transformed through some of the most recent solvers (particularly robust PCA, etc...). Stay tuned to http://nuit-blanche.blogspot.com/search/label/CAI
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.