Comparison of Huygens deconvolution with another well-known software package

Logically, this is a questions we are frequently confronted with which is why we have created this wiki article.

Ideally, one should try to get the best result out of each individual software package and use that for comparison. It does not make a lot of sense to compare similar number of iterations and other parameter settings, as the deconvolution approach taken varies widely among different softwares.

If your expectations of Huygens are not fulfilled by the image quality that Huygens produces, please contact us. We are ready to help you at any time with your image processing.

Below, we present a case study between Huygens and another software package "X".

We have compared images of HeLa cell in metaphase, deconvolved with Huygens and with another deconvolution package, named "X".

We thank Dr. Mikhail Bashkurov from the "Samuel Lunenfeld Research Institute" in Toronto (Canada) for providing these images. Channels represent HAUS6 (525 nm; green), pericentrin (605 nm; orange), and microtubules (705 nm; red).

After deconvolution with "X" the images look improved, but there are some issues that certainly need to be addressed.

Firstly, blurring remains present after deconvolution with "X". Below is a XZ projections of images deconvolved with the "X" (left) and Huygens (right). The left image shows clearly remnants of the blurring cone that have not been properly restored.

The blurring remnants can be observed as a diffraction pattern appearing as rings in the out of focus planes (see figure below and Image Formation).

Secondly, deconvolution with "X" shows centrioles in focus with dark halo's around these high intensity objects. This points to ringing artifacts which is a sign of over-restoration. Shown is one plane of a Z-stack.

Thirdly, the remaining background signal in the image after deconvolution with "X" is higher than in Huygens and the gain in signal after deconvolution is not as much improved as with Huygens. The latter suggests that the blurring is not optimally restored by deconvolution with "X".

You can create a line intensity profile (seen on the left) of images opened in the Twin Slicer by drawing a line with the mouse cursor, and by selecting under 'Plot', 'Plot both slicers'. When comparing the intensities between "X" (straight line) and Huygens (dashes line)' in a randomly taken area, you can see a clear difference in peak highs and lows (dynamic range), and the improvement in distinguishing detail. Image "X" is hidden by the graph on the left and Huygens is shown on the right.

Result after Huygens deconvolution.

Below you can observe the raw (left) and Huygens-deconvolved image (right). You can clearly see the improvement in contrast, and that more structural detail can be identified. Please do not compare the intensities as the brightness was adjusted to optimize the visibility of structural detail in both images.