search

starteam / starcellbio_html

2 stars 7 forks source link

Discovery: can we make noise more realisitic? #645

Open pdpinch opened 8 years ago

pdpinch commented 8 years ago

Currently the noise is applied to y values, and the amount of noise increases with y value. This makes the peaks look "realistic" but for low values of y the curve is still very smooth.

Suggestion: can we make the noise uniform across the y dimension, instead of having increase with the value of y?

pdpinch commented 8 years ago

Some samples from real research, collected by Lourdes:

https://www.dropbox.com/sh/t2vsaor8le2gwuq/AAArTkI0mOJbzUAzqNjVgTgba?dl=0

https://www.google.com/search?q=flow+cytometry&espv=2&biw=1436&bih=724&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiN-pa0yL7NAhUDPz4KHe90DRIQ_AUIBigB#tbm=isch&q=flow+cytometry+histograms&imgrc=eEIll8f6vSAmDM%3A

https://www.google.com/search?q=flow+cytometry&espv=2&biw=1436&bih=724&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiN-pa0yL7NAhUDPz4KHe90DRIQ_AUIBigB#tbm=isch&q=flow+cytometry+histograms&imgrc=Z75U5vZQIArkAM%3A

annagav commented 7 years ago

The noise needs to be more realistic with more spikes on the shoulders of the histogram. Maybe makes the data points more evenly spaced will help.