More Like this

1. Every Breath You Don’t Take: Deepfake Speech Detection Using Breath

   Layton, Seth; DeAndrade, Thiago; Olszewski, Daniel; Warren, Kevin; Butler, Kevin; Traynor, Patrick; Gates, Carrie (September 2025, ACM Digital Threats: Research and Practice (DTRAP))

   Deepfake speech represents a real and growing threat to systems and society. Many detectors have been created to aid in defense against speech deepfakes. While these detectors implement myriad methodologies, many rely on low-level fragments of the speech generation process. We hypothesize that breath, a higher-level part of speech, is a key component of natural speech and thus improper generation in deepfake speech is a performant discriminator. To evaluate this, we create a breath detector and leverage this against a custom dataset of online news article audio to discriminate between real/deepfake speech. Additionally, we make this custom dataset publicly available to facilitate comparison for future work. Applying our simple breath detector as a deepfake speech discriminator on in-the-wild samples allows for accurate classification (perfect 1.0 AUPRC and 0.0 EER on test data) across 33.6 hours of audio. We compare our model with the state-of-the-art SSL-wav2vec and Codecfake models and show that these complex deep learning model completely either fail to classify the same in-the-wild samples (0.72 AUPRC and 0.89 EER), or substantially lack in the computational and temporal performance compared to our methodology (37 seconds to predict a one minute sample with Codecfake vs. 0.3 seconds with our model) 

   more » « less
2. Adapt to warming and catch your breath

   https://doi.org/10.1038/s41564-018-0228-1  

   Hutchins, David A. (September 2018, Nature Microbiology)
3. Raman scattering applied to human breath analysis

   https://doi.org/10.1016/j.trac.2024.117791  

   Arachchige, Charuka Muktha; Muller, Andreas (August 2024, TrAC Trends in Analytical Chemistry)

   The chemical composition of exhaled human breath can be strongly correlated to medical conditions such as lung cancer or gastrointestinal diseases. To establish these correlations and, most importantly, to use them in diagnostics, chemical gas detection needs to be performed at trace concentrations, typically at parts-per-million (ppm) levels or below, for many compounds simultaneously. Traditional methods such as gas chromatography, a workhorse in scientific laboratories, is ill-suited for the fast, inexpensive point-of-care diagnostics that would be needed to build statistically-meaningful ensembles over large populations. With the increasing availability and decreasing cost of high power diode lasers and of uncooled CMOS cameras, spontaneous Raman spectroscopy (SRS), a vibrational molecular fingerprinting tool, is emerging as an economic alternative. Although gas SRS scattering cross sections are only on the order of 10$$^{-31}$$ cm$^2$/sr, considerable progress in the development of enhancement techniques has been made over the past decade. The purpose of this work is to review SRS enhancement approaches in the context of established human breath tests, and to provide a comparison with alternatives. Already, numerous trace gases such as H$$_2$$, CH$$_4$$, $$^{13}$$CO$$_2$$, and volatile organic compounds like acetone can be rapidly quantified in breath at concentrations below 1 ppm with SRS. With improvements in resolution and design of enhancement systems, SRS-based sensors could be scalably deployed in, e.g., pharmacies, and non-invasively screen for dozens of analytes at the parts-per-billion level. 

   more » « less
4. Reduced Graphene Oxide-Metalloporphyrin Sensors for Human Breath Screening

   https://doi.org/10.3390/app112311290  

   Lee, Bo Mi; Eetemadi, Ameen; Tagkopoulos, Ilias (December 2021, Applied Sciences)

   The objective of this study is to validate reduced graphene oxide (RGO)-based volatile organic compounds (VOC) sensors, assembled by simple and low-cost manufacturing, for the detection of disease-related VOCs in human breath using machine learning (ML) algorithms. RGO films were functionalized by four different metalloporphryins to assemble cross-sensitive chemiresistive sensors with different sensing properties. This work demonstrated how different ML algorithms affect the discrimination capabilities of RGO–based VOC sensors. In addition, an ML-based disease classifier was derived to discriminate healthy vs. unhealthy individuals based on breath sample data. The results show that our ML models could predict the presence of disease-related VOC compounds of interest with a minimum accuracy and F1-score of 91.7% and 83.3%, respectively, and discriminate chronic kidney disease breath with a high accuracy, 91.7%. 

   more » « less
5. Detection of Acetone on Human Breath Using Cyclic Voltammetry

   https://doi.org/10.1149/1.368441  

   Motsegood, P.; Leddy, J. (January 2012, ECS transactions)

   A cyclic voltammetric measurement protocol for acetone concentration collected in the vapor phase and measured in solution is demonstrated for acetone concentrations across the human physiological range, 1 μM to 10 mM at platinum electrodes in 0.5M H2SO4. Effects arise through adsorption of acetone from the gas phase onto a platinum surface and hydrogen in acidic solution within the voltammetric butterfly region. The protocol is demonstrated to yield breath acetone concentration on a human subject within the physiological range and consistent with ketone urine test strip. 

   more » « less