Technology
VOINOSIS establishes a technological advantage based on its proprietary
voice analysis AI technology and intellectual property.
Quantifying early signals through how people
speak, respond, and hear
VOINOSIS technology goes beyond basic speech recognition.
Our core innovation lies in AI biomarker technology that analyzes how people
speak, respond, and hear to identify early changes related to cognitive decline.
Core Technologies
1
Acoustic biomarker analysis
2
Hierarchical acoustic biomarker extraction
and inference pipeline
3
Speech-gaze multimodal ensemble technology
4
Language-agnostic analysis architecture
5
Pretrained speech encoder-based modeling
6
Digital assessment and training technologies
for auditory-cognitive screening
Intellectual Property & Technology Assets
Voinosis is building a comprehensive intellectual property portfolio spanning speech-based cognitive screening, auditory analysis, and AI inference architectures. Our technology assets cover core algorithms, multimodal analysis methods, disease screening models, and digital healthcare application frameworks—establishing a strong foundation for both defensibility and scalability in commercialization.
This technical capability has been validated on the international stage, including 1st place in the ICASSP MADRess Challenge 2023 and Joint 2nd Place at the 2023 Global Startup World Cup. Beyond product-level competitiveness, we pursue a sustainable IP strategy designed for global regulatory alignment and international expansion, advancing technology development and commercialization in parallel.
Intellectual Property & Technology Assets
Voinosis has fortified a world-class intellectual property portfolio encompassing speech-based cognitive
screening, auditory analysis, and advanced AI inference architectures.
Our assets extend beyond application interfaces to deep-tech core algorithms, multimodal analysis methods,
and latent space disease modeling, establishing an unassailable barrier to entry against global competitors.
Driven by a sustainable IP strategy tailored for global regulatory approvals (e.g., FDA) and international
expansion, we continuously accelerate our technology commercialization.
How We Listen
We Analyze How It Is Said —
Not Just What Is Said
Conventional speech AI relies on speech-to-text
transcription — inheriting its errors, language dependencies,
and fragility in real-world conditions.
Voinosis takes a fundamentally different approach. We capture the raw acoustics of the human voice: speech rate, intensity, micro-tremors in the vocal cords, and prosody. These are the most direct physical indicators of cognitive change — signals that exist before words, beyond language, and beneath conscious control. Our acoustic-first architecture ensures robust, consistent performance across languages, dialects, and noisy clinical or mobile environments.
Hierarchical Analysis
Task-Optimized AI, Built in Layers.
We do not rely on a flawed 'one-size-fits-all' model.
Our extraction pipeline is customized by clinical task.
Voinosis takes a fundamentally different approach. We capture the raw acoustics of the human voice: speech rate, intensity, micro-tremors in the vocal cords, and prosody. These are the most direct physical indicators of cognitive change — signals that exist before words, beyond language, and beneath conscious control. Our acoustic-first architecture ensures robust, consistent performance across languages, dialects, and noisy clinical or mobile environments.
Voice + Eye
Speech + Gaze, Cross-Verified Diagnosis
By fusing speech data with advanced eye-tracking signals,
we overcome the inherent limitations of single-modality systems.
Voinosis takes a fundamentally different approach. We capture the raw acoustics of the human voice: speech rate, intensity, micro-tremors in the vocal cords, and prosody. These are the most direct physical indicators of cognitive change — signals that exist before words, beyond language, and beneath conscious control. Our acoustic-first architecture ensures robust, consistent performance across languages, dialects, and noisy clinical or mobile environments.
From Diagnosis to Daily Life,
Built to Scale
From Clinical Diagnosis to
Daily Monitoring
A continuous, scalable loop bridging the gap between
hospital care and everyday life.
Voinosis thrives in the real world. BGD enables hospital-based cognitive screening and early detection, while HAHA Care extends proactive management and digital therapeutics directly into the patient's home. Together, they form a seamless care loop: detect early, monitor continuously, and intervene proactively. It is healthcare that does not end when the patient leaves the clinic.
A Foundation Engine Built for
Limitless Expansion
Disease-agnostic and language-agnostic architecture primed for
seamless global deployment.
Our proprietary speech encoder is not merely a disease-specific tool. It is a foundation model designed from the ground up to be independent of specific languages or single conditions. Beyond dementia, the architecture is engineered for immediate expansion into screening hearing impairment, depression, Parkinson's disease, and other neurological conditions — positioning Voinosis for rapid global scaling without the need for extensive retraining.
Why Voinosis
01
We analyze pure Acoustics,
not just text
While conventional speech AI relies heavily on STT transcription ("what is said"), Voinosis focuses on the fundamental acoustics ("how it is said"). Physical traits like speech rate, intensity, micro-tremors in the vocal cords, and prosody are the most direct evidence of cognitive shifts. This acoustic-first approach bypasses transcription errors and ensures unparalleled robustness in noisy environments (e.g., mobility) and across different languages.
02
We deploy task-optimized
Hierarchical Analysis
We do not rely on a flawed 'one-size-fits-all' model. Instead, we optimize our AI architecture by task level—from phoneme, to phrase, to full utterance—customizing the extraction pipeline for WRS, CBT, and PDT to capture the most precise, disease-relevant biomarkers.
03
We pioneer Multimodal
Screening (Speech + Gaze)
By fusing speech data with advanced eye-tracking signals, our platform overcomes the limitations of single-modality systems. This ensures high diagnostic accuracy even for patients with language barriers or illiteracy, providing clinicians with the most reliable, cross-verified screening results.
04
We bridge the gap between
clinical care and daily life
Our technology thrives in the real world. By deploying BGD in hospitals and HAHA Care in homes, Voinosis has commercialized a continuous, scalable healthcare ecosystem that spans from early detection and monitoring to proactive management and digital therapeutics.
05
Our core engine is built for
limitless expansion
Our proprietary speech encoder is a foundation model designed to be both disease-agnostic and language-agnostic. Beyond dementia, our architecture is primed for immediate expansion into screening hearing impairment, depression, Parkinson's, and other neurological conditions, positioning us for seamless global scaling.
