Author Archives: Rich

Gimp ‘dust and scratches’ filter

Posted on by
3

Photoshop (and Elements) has a terrific filter called Dust and Scratches. It looks for dots and curves on the image that are clearly noise, then removes them. This comes from dust on your lenses or on the photos you’re scanning, etc. An invaluable tool for all image cleanup work.

No such thing in the Gimp world. Search for gimp + dust + scratches and you either find the despeckle filter (which is useless, ime), or references to some 3rd-party plugin which was never compiled for windows.

The best tool I know of comes from the G’MIC plugin for Gimp. This is a must-have tool for many reasons (e.g., making lady’s portrait skin look soft and smooth), but the Remove Hot pixels option under the Repair subset is essentially the same thing as the Dust and Scratches of Photoshop.

I have not read the technical documentation, but the controls are simple: Mask size is the number of pixels at which to consider something being a “hot pixel”. Too low and you miss pixels. Too high and you lose detail in the image. Threshold is the amount of difference to require for something to be considered “hot”. Too high and you won’t change anything. Too low and you change too much. A standard approach is to start with mask size and threshold at minimum, then increase mask size until you fix all/most of the errors, then increase threshold until you stop screwing up everything else. Tip: As this is a compromise, leave the big scratches for manual work: Use the Clone and Healing tools instead.

Here is an example of a BEFORE ==> AFTER.

UPDATE: January 2016

As of the latest G’MIC plugin for Gimp, this effect has been moved to “Repair -> Remove hot pixels”

Also, try the “Repair -> Iain’s noise reduction”

Some Experiments on Repeatability of Wolfram’s Rule 30 (Written: 10/22/2004)

Posted on by
Reply

For my first CA programming task with Java, I decided to implement Wolfram’s rule 30. This led to some questions, i.e., just how wide should the cell row be, and what to do about the endpoints. When I started the work I could not find Wolfram’s discussion on the matter. (Yes I have his book and yes I have it online as well; I just missed it.)  So I ran some studies myself. I discovered that the repetition period is actually very low. I was shocked at how low it is. My writeup can be found at  expts_with_rule30

Rovarii meet Obama, 2008 Photo Album

Posted on by
Reply

11/1/2008 Hello all, Yesterday Chris, Rachel and I went to the Obama rally in Pueblo (30 miles south of the Springs) and had a most pleasant surprise. We left the really a little early to get a table at what we were told is “the best local Mexican food restaurant” in the city. We just got seated after a 20 minute wait when a staff member came to our table and told me that he liked the Obama pin on my hat and that the Obamas were on their way to have dinner. We had a seat with a picture window that faced the street and within minutes the Secret Service and police showed up, shut down the street and restaurant. A few minutes later a motorcade with the Obama’s car stopped in front of our window and the Obamas came in with their children. Barrack stopped at our table and shook everyone’s hand and then joined his family a few feet away. Later Michele Obama came by and I got to “bump fist” with our next First Lady. Chris was impressed by the security and said it was like being on the set of the Matrix. Rachel was ready to follow the advice of her cousin Jaia and ask the Obamas to pass the Ketchup, but being in a Mexican restaurant they did not have any on the table to pass. Here are some of our photos from the restaurant. Enjoy. Mike

Solvista, 2007 & 2008 Photo Album

Posted on by
Reply

Pearl and the kids ski for the first time. Solvista is undoubtedly the best deal in Colorado for learning to ski. Very family-friendly. Great instructors. Low prices. Highly recommended for first and second timers.

[shashin type=”album” id=”19″ size=”max” crop=”n” columns=”2″ caption=”y” position=”center”]