Kardome was founded in 2019 by Dani Cherkassky and Alon Slapak in Tel Aviv, Israel, with a mission to perfect speech recognition and voice communication applications. Frustrated with existing voice command and speech recognition devices the two founders combined their extensive knowledge in acoustics, advanced machine learning, and signal processing algorithms to create Kardome’s VUI technology, offering a solution that makes individual voices understandable for speech recognition engines and human listeners.
Now, Kardome has expanded its offerings by adding voice biometrics and wake word detection. The addition of the two technologies extends the company's ability to help developers, manufacturers, and OEMs enhance their existing automatic speech recognition (ASR) systems. Combined with Kardome's 3D Audio Front End software, the new voice products provide a holistic solution for edge software on all voice-enabled devices.
Kardome’s 3D Audio Front End technology uses location-based, source separation, and noise reduction algorithms to improve ASR system performance substantially. Adding Kardome Wake and Voice ID further personalizes a secure voice recognition experience. Essentially, Kardome enables product developers to know who is talking, where they are, and what they’re saying.
Voice biometrics, or Voice ID, is used to verify a person's identity based on their unique voice characteristics, such as pitch and pronunciation. Wake word detection triggers the ASR system using a user's custom verbal command. Kardome's voice biometrics and wake word detection software offer OEMs an easy way to add these capabilities to their existing devices without the cost and time of assembling off-the-shelf edge software, disjointed building blocks, or needing a connection to the cloud.
Kardome Voice ID enables developers and OEMs to build a voice user interface that recognizes users and responds only to them. The company's voice biometric technology provides security and accurate identification of who is speaking. Users need only to create a voice profile; from that moment, Kardome recognizes them even if they move around the room while they speak. Kardome Voice ID is always on, ready to detect who is speaking.
On its own, Kardome's Voice ID allows developers and manufacturers to add voice authentication technology to their applications, enabling users secure access to a device or data. Kardome’s deep learning and noise reduction algorithms adds a highly accurate Wake Word detection and seamlessly integration with our the company's Echo Cancellation and Noise Reduction solution.
Kardome claims its Wake Word technology offers the highest recognition accuracy, with a False Response Rate of less than 10% with an SNR of -20db. Kardome is always on, ready to detect custom wake words, and works with embedded systems, offering secure and fast response times without requiring Internet connectivity. Kardome says it is able to support development on edge devices working with any ASR system, based on Qualcomm, HIFI4, ARM, ADSP-SC58x/59x, and more.
www.kardome.com
