We’ve been reviewing the output coming from the full Planet Four data reduction pipeline now that we’ve frozen development on the fan and blotch clustering codes. Once we have the fan markings and blotch markings clustered individually, we then have a stage that combines the individual clusters to decide it a source marked by Planet Four volunteers is really a blotch or fan by find clusters where there centers on top of each other and then depending on how many fan markings went into the fan cluster and how many markings went into the blotch cluster we decide it’s a fan or a blotch for the final catalog. What we found in the catalog review that there are nice cases where there are sources that aren’t quite a fan only or not just a blotch. With a citizen science approach we’re able to capture that fuzziness which is fantastic. We highlight a few examples below selected by Michael Aye, who has been hard at work developing this pipeline over the past several years.
In all three figures: the top left is the Planet Four subject image, top middle is all the individual volunteer fan markings, top right is all the individual volunteer blotch markings, the bottom right is the blotch clusters after clustering, the middle bottom is the fan clusters after clustering, and the bottom left is the final sources after combining the fans and blotches (the dots in this panel show the center position of the final fan or blotch in our catalog).
As you can see from above, we’re making great progress on the Planet Four data reduction pipeline. Next steps including handing the fact that the edges of most of our Planet Four subject images overlap with neighboring subject images, and ensuring that we merge overlapping volunteer markings covering the same spot on two different subject images.
Happy Birthday Planet Four: Terrains! Last Thursday marked the second anniversary of the launch of Planet Four: Terrains. Thank you for all of your help over the past two years. We couldn’t do this without you. To celebrate here’s a recipe for a Martian South Polar Trifle for you to make for your own Mars party. It’s inspired by this recipe:
- 1 package (20 ounce) chocolate cream filled cookies like Oreos or Hydrox
- 16 ounces of cream cheese
- 1/4 cup of unsalted butter
- 2 cups of powdered sugar
- 2 small packages of instant vanilla pudding mix
- 3.5 cups milk (recommend dairy milk – skim to whole works fine)
- 1 tub (20 ounce) of whipped topping (not canned whipped topping but the kind you find in the freezer section). ( You can also make your own from scratch using whipping cream if you like. Just add a little bit of powdered sugar mixed in to taste)
Making the ground layer: Place chocolate cream filled cookie in a bowl or plastic bag, saving 4-6 cookies for the top later. Use a rolling pin, bottom of a glass, etc to crush the cookies. Place 3/4 of the crushed cookies on the bottom of a nine-inch glass pie plate or any glass container so you can see through the sides if possible.
Making the seasonal ice sheet mixed with seasonal ice and boulder: Soften the butter., then mix together cream cheese, butter, and powdered sugar until smooth (electric mixer or whisk is fine). With whisk combine the pudding mix with the milk until smooth and slightly thickened. Fold together cream cheese mixture, whip topping, and pudding mixture. Add 2-3 drops of red food coloring to get a light red color (the red food color and the power sugar represent the atmospheric dust trapped in the ice sheet). Then mix in the remaining 1/4 of the crushed cookies to represent boulders trapped in the ice sheet. Pour the mixture over the ground layer of cookies.
Topping of seasonal fans: Crush the remaining few cookies very finely, sprinkle on the top to create seasonal fan shapes
Refrigerate the trifle for 3 hours or more before serving.
I’ve been reviewing output from the blotch clustering with updated parameters, and the good news is the two problematic cases we were trying to have the clustering algorithm identify now are successful:
We’re looking a bunch of parameters, but the bottom left corner configuration is the parameters we think are best for the blotches and and fans.
Here we are showing just the clustering of the fan markings.
Here we are showing just the clustering of the blotch markings
Here we show the clustering of the blotch markings. Now the 2-scale blotch clustering identifies the volunteer draft markings for the large blotch below as well as smaller ones.
I think we’re at the point of locking development of the blotch pipeline. This will be a big step forward towards finishing the first paper and getting science results out of Planet Four. We’ll want to take a look at the next stage of the pipeline after two different types of markings are cluster and the algorithm picks the final shapes based on the number of classifiers who drew fans versus blotches for the same source. If that looks good as we expect it should from previous reviews of the pipeline output, then the next thing we need to do is look at the pipeline results near the edges of each of the cutouts where this is overlap between the different subject images.