source : www.miragenews.com
Researchers have developed a new deep learning AI tool that generates lifelike bird songs to train bird identification tools, helping ecologists monitor rare species in the wild.
Identifying common bird species by their songs has never been easier, with numerous phone apps and software available to both ecologists and the public. But what if the identification software has never heard a particular bird, or only has a small number of recordings for reference? This is a problem ecologists and conservationists face when monitoring some of the world’s rarest birds.
To overcome this problem, researchers from the University of Moncton, Canada, have developed ECOGEN, a first-of-its-kind deep learning AI tool, which can generate lifelike bird sounds to improve samples of underrepresented species. These can then be used to train audio identification tools used in ecological monitoring, which often contain disproportionately more information about common species.
The researchers found that adding artificial bird song samples generated by ECOGEN to a bird song identification improved bird song classification accuracy by an average of 12%. The findings were published in Methods in ecology and evolution.
Dr. Nicolas Lecomte, one of the lead researchers, said: “Due to significant global changes in animal populations, there is an urgent need for automated tools, such as acoustic monitoring, to track shifts in biodiversity. The AI models used to identify species in the field of acoustic monitoring lack extensive reference libraries.
“ECOGEN allows you to close this gap by creating new instances of bird sounds to support AI models. Essentially, for species with limited wild recordings, such as those that are rare, elusive or sensitive, you can expand your sound library without further steps to disrupt of the animals or carrying out additional field work.”
An AI tool for conservation
The researchers say that creating synthetic bird songs in this way could contribute to the conservation of endangered bird species and also provide valuable insight into their vocalizations, behavior and habitat preferences.
The ECOGEN AI tool has other potential applications. For example, it could be used to help conserve extremely rare species, such as the critically endangered regent honeyeaters, where young individuals cannot learn the songs of their species because there are not enough adult birds to learn from.
The tool could also benefit other species. Dr. Lecomte added: “Although ECOGEN was developed for birds, we are confident it can be applied to mammals, fish (yes, they can produce sounds!), insects and amphibians.”
In addition to its versatility, an important advantage of the ECOGEN tool is its accessibility, as it is open source and can be used even on basic computers.
How does it work?
ECOGEN works by converting real recordings of birdsong into spectrograms (visual representations of sounds) and then generating new AI images from these to expand the dataset for rare species with few recordings. These spectrograms are then converted back into audio to train bird sound identifiers. In this study, the researchers used a dataset of 23,784 recordings of wild birds from around the world, including 264 species.
/Public publication. This material from the original organization/author(s) may be contemporary in nature and has been edited for clarity, style and length. Mirage.News takes no institutional positions or parties, and all views, opinions and conclusions expressed herein are solely those of the author(s). View the full document here.
source : www.miragenews.com