Matthias Arzt

CachedOpEnvironment returns wrong results due to hash collision

4

I created a test that illustrates this problem: see https://github.com/maarzt/imagej-ops/commit/791a1b81b1962ab1a5ce0f9221d5ff76dcd32be6

Wrongly typed signature of ops.stats().mean(Img<Float>)

2

The following piece of code produces a ClassCastException on runtime: ``` Img img = ArrayImgs.floats(1, 1); FloatType mean = ops.stats().mean(img); ``` The method `ops.stats().mean(Img)` always returns an instance of DoubleType...

CommandService.run(...).get() crashes when used in UI-Thread

12

Many new ImageJ2 / FIJI developers run into this bug. So we should fix it soon: `CommandService.run(...).get()` blocks, when used in the EDT (UI-Thread): ```java final Future< CommandModule > future...

Fix Logger Parameter with Ops

6

Previously the LoggerPreprocessor automatically injected Logger Paramters for CommandService. But LoggerPreprocessor didn't work for Ops. This PR fixes this issue: 1. Context injects Logger Parameter 2. Logger extend Optional interface,...

The way Labkit uses resources to batch process images can still be improved: * process small images in parallel * set optimal TaskExecutors for CPU and GPU processing * match...

Currently a user has to select "use gpu" every time when opening a classifier or adding a new one to the list. This should be changed, the UI should remember...

Concepts for labeling in 3d

1

If you think of software like Blender it becomes clear, that editing a 3d volume is a very difficult task, which requires versatile tools. But which functional, that we can...

1. Labkit only works with the first few but shows all. 2. If the affine transformation for every time frame, then there's an offset between image and segmentation.