1
$\begingroup$

Usually, in MR images one hypothesis research communities are making is, " the spatial intensity distribution is piecewise constant and that each tissue corresponds to a unique grayscale level. Based on that hypothesis, a valid correction method should lessen the standard deviation in intensity for each tissue."(Paper below)

Belaroussi, B., Milles, J., Carme, S., Zhu, Y. M., & Benoit-Cattin, H. (2006). Intensity non-uniformity correction in MRI: existing methods and their validation. Medical Image Analysis, 10(2), 234-246.

Now, in case of underwater images, is THIS hypothesis valid considering the fact that artificial lighting conditions, ocean floors and objects as well as object colours may vary in any two subsequent frames?

What about the in-class variances, before and after the correction of bias/illumination?

Image courtesy: Woods and Hole Oceanographic Institute

Image courtesy: Woods and Hole Oceanographic Institute

Image courtesy: Woods and Hole Oceanographic Institute

For more images: habcam.whoi.edu/gallery/

| cite | improve this question | |
$\endgroup$
  • 1
    $\begingroup$ I'm not sure that this interesting question is quite on-topic here. $\endgroup$ – Yuval Filmus Mar 9 '17 at 15:52
  • $\begingroup$ I am not sure if it is on-topic either. But as per my opinion, there is no way anyone can define the range of intensities for a particular object. Because it may vary for number of reasons as mentioned above. Even if someone come up with a unique intensity range if that object is not there in 100/120 images hows that range will help you to validate the bias correction. On second question about the in-class variance, I am totally blank. $\endgroup$ – Jay Patel Mar 9 '17 at 16:07
  • $\begingroup$ Can you please articulate the hypothesis, in the context of your ocean floor images, more precisely? Are you talking about video? Are you talking about consecutive frames of the video? What are your thoughts? If your question is "Are object intensities constant after illumination correction, considering that illumination may vary?", then presumably you know the answer to that question. If your question is "re object intensities constant after illumination correction, considering that objects may vary?", why would you possibly expect the answer to be yes? $\endgroup$ – D.W.♦ Mar 9 '17 at 18:50
  • $\begingroup$ I have added the hypothesis about the MR data. Second, I am talking about the images, not videos.Usually, when any institute takes the ocean floor images they take overlapping images which ultimately creates the whole ocean floor. Same as PANORAMA. Third, if I take a hypothesis that scallops for ex in the images shown refer to a unique grayscale intensities, but Illumination, scallops colors (brown, reddish brown, light brown, pink brown, white, off-white, yellowish white, scallops having sand particles on its top) varies greatly. Will it refer to unique grayscale after correction? $\endgroup$ – Jay Patel Mar 10 '17 at 3:36
  • $\begingroup$ About in-class variance, with-in tissue, intensity variance should decrease after correction in MR data. What about the same thing if I consider for Ocean images. Eg. if I take a manually segmented background and measure before and after variance of that segmented patch, should variance be decrease in this case too? $\endgroup$ – Jay Patel Mar 10 '17 at 3:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.