Interest point detection in Descriptor based (2d/3d) works but not in Bigstitcher

[yamalcaraz]

Hi,

First of all, I am a fan of your work and it is very helpful in my research.

Issue
When I use the Descriptor based (2d/3d) application I could successfully align the image below.

However, when I try it with bigstitcher, I can go to the point where i define the sigmas and the threshold as seen in following two images.

After pressing done, I don't get any interest points at all, and I get the following log:

  (Fri Apr 29 21:41:48 CEST 2022): Subpixel localization using quadratic n-dimensional fit
  (Fri Apr 29 21:41:48 CEST 2022): Warning! Failed to compute subpixel localization SubpixelLocalization: [List<DifferenceOfGaussianPeak<T>> peaks] is empty.
  (Fri Apr 29 21:41:48 CEST 2022): Found 0 peaks.
  (Fri Apr 29 21:41:48 CEST 2022): Correcting coordinates for downsampling using AffineTransform: 3d-affine: (1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0)
  (Fri Apr 29 21:41:48 CEST 2022): WARNING: List is empty.
  (Fri Apr 29 21:41:48 CEST 2022): Requesting Img from ImgLoader (tp=0, setup=9), downsampling: (1, 1, 1)
  (Fri Apr 29 21:41:48 CEST 2022): Using precomputed Multiresolution Images [1x1x1], Remaining downsampling [1x1x1]
  (Fri Apr 29 21:41:49 CEST 2022): min intensity = 0.0, max intensity = 255.0
  (Fri Apr 29 21:41:49 CEST 2022): computing DoG with (sigma=1.8015, threshold=0.25585005, sigma1=(1.7307230234146118, 1.7307230234146118, 1.7307230234146118), sigma2=(2.0831925868988037, 2.0831925868988037, 2.0831925868988037))
  (Fri Apr 29 21:41:49 CEST 2022): Detecting peaks.
  (Fri Apr 29 21:41:49 CEST 2022): Mem-copying image.
  (Fri Apr 29 21:41:49 CEST 2022): Quadratic localization.
  (Fri Apr 29 21:41:49 CEST 2022): Subpixel localization using quadratic n-dimensional fit
  (Fri Apr 29 21:41:49 CEST 2022): Warning! Failed to compute subpixel localization SubpixelLocalization: [List<DifferenceOfGaussianPeak<T>> peaks] is empty.
  (Fri Apr 29 21:41:49 CEST 2022): Found 0 peaks.
  (Fri Apr 29 21:41:49 CEST 2022): Correcting coordinates for downsampling using AffineTransform: 3d-affine: (1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0)
  (Fri Apr 29 21:41:49 CEST 2022): WARNING: List is empty.
  (Fri Apr 29 21:41:49 CEST 2022): DONE.

How do I resolve this? I was hoping to further align (rigid) several tiles together after stitching, that is why I am trying it in bigstitcher.

I hope you can help.

Many thanks!

[afshindini]

What is the solution for the above problem as I faced the same issue. Cannot detect any interest point in 2D images?!

(Mon Oct 21 09:31:58 EEST 2024): Subpixel localization using quadratic n-dimensional fit
(Mon Oct 21 09:31:58 EEST 2024): Found 0 final peaks.
(Mon Oct 21 09:31:58 EEST 2024): Correcting coordinates for downsampling using AffineTransform: 3d-affine: (1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0)
(Mon Oct 21 09:31:58 EEST 2024): WARNING: List is empty.
(Mon Oct 21 09:31:59 EEST 2024): min intensity = 0.0, max intensity = 255.0
(Mon Oct 21 09:31:59 EEST 2024): computing DoG with (sigma=1.8, threshold=0.008, sigma1=(1.7291616201400757, 1.7291616201400757, 1.7291616201400757), sigma2=(2.0813581943511963, 2.0813581943511963, 2.0813581943511963))
(Mon Oct 21 09:31:59 EEST 2024): Detecting peaks.
(Mon Oct 21 09:31:59 EEST 2024): Found 0 initial peaks (before refinement).
(Mon Oct 21 09:31:59 EEST 2024): Quadratic localization.
(Mon Oct 21 09:31:59 EEST 2024): Subpixel localization using quadratic n-dimensional fit
(Mon Oct 21 09:31:59 EEST 2024): Found 0 final peaks.
(Mon Oct 21 09:31:59 EEST 2024): Correcting coordinates for downsampling using AffineTransform: 3d-affine: (1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0)
(Mon Oct 21 09:31:59 EEST 2024): WARNING: List is empty.
(Mon Oct 21 09:31:59 EEST 2024): DONE.