Aleksandr Laptev

Title: Innovations by Aleksandr Laptev in Audio Processing

Introduction

Aleksandr Laptev is an inventor based in Erevan, Armenia. He is known for his contributions to audio processing technologies, particularly in multi-speaker and multi-channel environments. His work focuses on utilizing machine learning techniques to enhance speaker recognition and verification.

Latest Patent Applications

Aleksandr Laptev has filed several notable patent applications. One of his latest applications is titled "AUDIO PROCESSING IN MULTI-SPEAKER MULTI-CHANNEL AUDIO ENVIRONMENTS." This application discloses apparatuses, systems, and techniques that may use machine learning for implementing speaker recognition, verification, and/or diarization. The techniques include receiving a first set of audio data channels (ADCs) that jointly capture speech produced by one or more speakers. It also involves obtaining a second set of ADCs based on the first set, where individual ADCs represent channels of the first set. The processing of these ADCs is done using an audio processing neural network model to associate the speech with the respective speakers.

Another significant application is "PROCESSING OF AUDIO DATA IN MULTI-SPEAKER MULTI-CHANNEL ENVIRONMENTS." This application also utilizes machine learning for speaker recognition and verification. The techniques include processing audio data channels using a voice detection model to determine voice activity likelihoods (VALs). It further involves generating embeddings associated with the ADCs and processing these embeddings to obtain associations of the speech to the speakers.

Conclusion

Aleksandr Laptev's innovative work in audio processing showcases the potential of machine learning in enhancing communication technologies. His latest patent applications reflect a commitment to advancing the field of speaker recognition and verification.