Related publications: [submitted]
RAFAR is an algorithm designed for long-term analysis of foraging behavior. It uses the regularity of this behavior to recognize grazing and rumination bouts. Acoustic signals are analyzed in two main stages: segmentation (Seg) and classification (Cla).
Several variants can be evaluated by combining the main stages of the algorithm with additional ones. They differ in the order of stages in the process flow. Additional stages are gap merging before classification (MB), partitioning long blocks (BP), and gap merging after classification (MA).
This web-demo requires a file with the audio recording (an example is provided for download). The input sound must be a one channel (mono, 16-bit, 44.1 kHz) WAV file up to 15 min in length. After choosing the file, select the desired variation of the algorithm and click Submit. If no file is provided, example file will be loaded by default. The output is a TXT file with the timestamps and labels corresponding to the file analyzed. Input and output files can be imported in Audacity to get a graphical view of the recognition (click here for an example).
Contact: Sebastián Vanrell