MLE versus ICTM - Why is each method more effective under certain circumstances.
I was somewhat unsure about the difference between the MLE and the ICTM reconstruction techniques. I understand that both differ in the calculation of the difference measure and that the ICTM uses the least squares rule while the MLE minimizes the I-divergence. However, I don't quite comprehend the specifics and why each method is more effective under certain circumstances.
Not very easy...
It can be rigorously proven that when dealing with non-negative objects (fluorescing objects) the I-divergence criterion is the only consistent choice, whereas for objects which can be both positive and negative (e.g. sound) least squares are is the best choice. Another viewpoint is that I-divergence incorporates the Poisson nature of the emitted fluorescence light whereas least squares incorporates Gaussian noise. With this in mind one could expect that for low noise levels where the differences in the Poisson and Gaussian distributions are small there is no preference between the methods other than their computational efficiency. One reason why we still have a preference for the MLE is that it handles noisy backgrounds much better than the ICTM. A disadvantage of the MLE is that it easily overemphasizes small structures, but we constrain this.
In the Huygens Pro version prior to 2.3 the ICTM is a good deal more efficient than the MLE. Starting from version 2.3 the QML, based on a conjugate method for the MLE class algorithms the iterations are 2 to 5 times as effective as the former Fast-MLE method.Also the classic-MLE, which replaces the former MLE, is now only slightly slower than the ICTM.
See also: Mle Ictm Snr Imp Faq.
<br>
Keywords: MLE ICTM I-divergence Poisson Gaussian noise<br>
Categories: Faq Deconvolution, Huygens Faq, Imported Faqs<br>
Platforms: Irix Linux<br>
Related products: Hu Pro<br>
