F-POD software and user guides

The current version of the FPOD app, dated 20 Dec 2024.  It uses Version 1.0 of the advanced KERNO-F classifier.

The latest additions include an implementation of the PYRA method of estimating trends in detection rates / access to smoothed inter-click-interval values / and an export utility for users who want to develop their own classifiers using machine learning or AI.

Updated: December 2024

The F-POD app includes a lot of display, filtering and exporting options.

Hopefully this guide will be useful in learning your way round this beast!

Updated: November 2024

Data on SD cards is in .CHE files, which may be one per day or other configurations, but in all cases you read them with the FPOD app and they are then on the PC as .FP1 files.

Mac users will need a PC-emulator

Franciscana and Guiana dolphins recorded by Renan Paitach using a C-POD and an F-POD in Babitonga Bay in Brazil.

View these together in the FPOD app to see the differences between the instruments and species.

Excel template for adding annotations to          F-POD and C-POD files.

If you make observations, or arrive at species classifications using your own classifier, you can rapidly add your findings to the file so that the app will show your species designation on a minute by minute basis.  Then you can filter the data by the species information you added.

Updated: July 2024

Databases can be quicker and easier to use for large data sets such as those from POD studies.

This is a brief guide to using queries in Microsoft’s Access.

This is a reference document on software changes and is of limited interest!

The KERNO-F classifier that underpins the usefulness of F-PODs is based on deep ‘feature engineering’ to seek ‘coherent’ click trains and then allocates them to species guilds or leaves then ‘unclassed’.

This takes you to a paper on the validation of the results.

This takes you to a paper on PYRA.

Updated: December 2024

An interface now exists to allow you to extract detailed click, train and encounter features that you can use to develop AI or other classifiers.

Then you can inject your classification back into the FP3 file and see exactly how well it did!