Detrending is a technique to remove unwanted trends from time-series data. Image series (videos) may be viewed as a collection of time series: each pixel is its own time series, the value at time t being the intensity value of that pixel in the frame recorded at time t. Detrending is applied to image series in the fields of fluorescence fluctuation and correlation spectroscopy (FCS and FFS) to remove trends introduced by photobleaching and also other possible sources of trends such as laser power fluctuation. detrendr is an R package for detrending image series.

If you’re new to R and you’re here because you want to use detrendr, be warned that you will need to learn some basic R first. I recommend reading the short book “Hands On Programming with R” by Grolemund. This is available for free at https://rstudio-education.github.io/hopr/. That should be enough but if you want further reading, check out “R for Data Science” which is available for free at https://r4ds.had.co.nz/.

This website gives an introduction to the detrendr package, assuming that the reader has a basic level of R knowledge.

## Installation

You can install the release version of detrendr from CRAN with:

install.packages("detrendr")

You can install the (unstable) development version of detrendr from GitHub with:

devtools::install_github("rorynolan/detrendr")

I highly recommend using the release version. The dev version is just for the ultra-curious and should be thought of as unreliable.

## Using detrendr

There are two ways to use detrendr.

1. Interactively in the R session, playing with the image as a numeric array, dealing with one image at a time.
2. In batch mode, having the software read TIFFs, perform the detrending and then write the detrended TIFFs to disk when detrending is over. This method permits the user to use R as little as possible and is better for those who don’t intend to become bona fide R users.

These are discussed in two articles.

### Linescan data

The article Linescan data shows how to deal with data in linescan (as opposed to stack) format. If you don’t know what linescan data is, you don’t need to read this article.