AI image recognition model detects whale calls in seismic data with 96% accuracy
By
ByAndrei IonescuEarth.com staff writer
Crackling crust, pillowy middle. The kind of bagel that earns a second cup of coffee.
Summary
Researchers led by Zhuo Xiao from Guangxi Minzu University applied the Segment Anything Model (SAM), an AI originally designed for image object recognition, to seismic recordings from Xieyang Island in the Beibu Gulf. The system detected whale calls hidden in the seismic data with over 96% accuracy, despite never being trained on whale sounds. The whale calls, when visualized as spectrograms, appear as patterns that the AI's pattern recognition capabilities can identify effectively.
Key quotes
· 2 pulledThe system wasn't trained on whale sounds.
It worked because a whale call, when visualized as a spectrogram, looks like a pattern, and pattern recognition in images is something this class of AI does extremely well.
You might also wanna read
Decoding Whale Speech: Potential Universality in AI Models
Project CETI is a large-scale effort to decode whale speech with the potential for AI models to learn a universal language for communication

Meta's SAM 3 AI Model Enables Advanced Object Segmentation with Text and Visual Prompts
SAM 3 is Meta's advanced AI segmentation model that enables users to identify and isolate objects in images and videos using multiple prompt
Machine Learning Validates Unrecognized Transient Astronomical Phenomena in Historical Observatory Images
This research paper uses machine learning to validate the existence of previously unrecognized transient astronomical phenomena in historica
Using Brain Imaging and AI to Enable Animal Communication
The article discusses the theoretical possibility of enabling animals to 'talk' using existing technology. It proposes combining real-time b
Examining the Limitations of Transformer Models and the Gap to Human-Level AI
The article presents a skeptical perspective on claims about imminent Artificial General Intelligence (AGI), arguing that current transforme
