Incorrect implementation of Christov2004 algorithm

[mh105]

I came across this repo looking for implementations of the Christov 2004 ECG detection method.

I don't think the code here is doing what the paper originally described. For example in

py-ecg-detectors/ecgdetectors.py

Line 231 in f39354c

F = F + ((max_latest-max_earliest)/150.0)

the integrating threshold F will grow indefinitely here - it should be F = mean(F_section) + ((max_latest-max_earliest)/150.0).

The way R decays slower is also wrong - it should be decaying 1.4 times slower not 1.4 times less.

Just raising a flag in case people try this and get sub-optimal results - these issues might be why.