VKG Analyzer - future of vocal diagnostics
VKG 3.0 integrates advanced videokymography with AI-powered analysis to deliver quatitative voice diagnostics.
VKG 3.0 integrates advanced videokymography with AI-powered analysis to deliver quatitative voice diagnostics.
VKG 3.0 is a modular software developed to analyze and interpret multi-line video kymographicdata, aiming to assist medical professionals in diagnosing vocal fold disorders. The project is a collaboration between UTIA AV ČR, ATE Systems, and the Hlasové Centrum Praha, financed by the Technology Agency of the Czech Republic (TACR), under grant number TH04010422.Our mission is to improve and make more quantitative vocal diagnostics with innovative software solutions.
Overview of the VKG 3.0 platform and its modular components.
The VKG Analyzer is designed for advanced analysis and visualization of videokymograms, enabling detailed examination of vocal fold vibrations.The module includes preprocessing tools to enhance data quality, such as noise reduction and smoothing. Advanced visualization features including color-coded outputs, highlight key patterns and abnormalities. Quantitative analysis capabilities measure critical parameters like closure duration and the ratio of closing-to-opening phases. It provides intuitive tools for clinicians.
The VKG Audio module integrates audio analysis with videokymographic data to provide a comprehensive view of vocal fold function. The module calculates essential acoustic parameters, such as sound pressure level, fundamental frequency, and pitch strength. The VKG Audio module helps identify voice disorders by analyzing sound irregularities linked to structural or functional vocal fold issues.
The VKG Simulator module simulates vocal fold vibrations. It generates high-speed simulations of vocal fold motion based on a kinematic model, enabling users to study both healthy and pathological conditions. The module allows adjustments to parameters such as frequency, amplitude, and mucosal wave propagation, offering realistic representations of different vibration patterns. This tool serves as a research and an educational resource, helping clinicians and researchers visualize and analyze intricate vocal fold mechanics.
The VKG Convertor module enables seamless conversion of legacy kymographic data into modern formats compatible with current diagnostic tools. It processes data from older VKG systems. This module is essential for expanding diagnostic capabilities by integrating past and present data into a unified platform.
The VKG Estimator module utilizes advanced machine learning AI techniques to automatically estimate critical vocal fold parameters - the sharpness of lateral peaks and the length of mucosal waves. It provides estimations, aiding in the early detection of vocal fold disorders and abnormalities. This module is a key innovation in VKG 3.0, offering automated analysis to enhance diagnostic accuracy and efficiency for clinicians.
The VKG 3.0 research is focused on advancing videokymography through AI-powered automation. The team developed methods to detect lateral peak sharpness and mucosal wave length using convolutional neural networks. A robust database of over 5,000 annotated video and audio recordings was created for diagnostic and research purposes. The research included validating algorithms with simulations and real clinical data, enhancing diagnostic accuracy. Findings were shared through international publications, conferences, and workshops, contributing to vocal health research.
The VKG 3.0 project was a collaborative effort involving experts from three fields: UTIA AV ČR: A leading research institute specializing in image and video processing and medical imaging, responsible for developing and implementing machine learning algorithms and data processing. Hlasové Centrum Praha: A clinical institution focused on diagnosing and treating voice disorders, providing medical expertise, data collection, and validation of diagnostic methods. ATE Systems: A private company with extensive experience in camera systems.
Explore our resources to learn more about VKG 3.0:
