CI Tech Unveils Noisy Environment Voice Recognition AI Technology at ICASSP 2026

CI Tech has demonstrated its world-class technological competitiveness by unveiling AI technology capable of accurate voice recognition even in noisy environments at the world’s most prestigious academic conference.

On May 8, CI Tech announced that it officially presented its self-developed voice recognition AI technology at ‘ICASSP 2026 (IEEE International Conference on Acoustics, Speech, and Signal Processing),’ an international conference in the field of signal processing. The company previously won first place with this technology at a global audio AI competition held in December last year.

This technology features the application of a multi-voice recognition model structure called ‘Stereo Chorus of Whispers.’ Multiple voice recognition models simultaneously analyze the same audio source and then combine left and right listening signals to mimic the way humans actually perceive sound.

According to CI Tech, prediction accuracy continues to improve as stereo processing, perceptual audio quality factors, and sentence length information are incorporated step by step. In addition, the technology’s scalable architecture—unlike single-model systems—enables it to be applied to a wide range of environments, which is also considered a major advantage.

This technology was recognized as the best in the field by winning first place among 24 competing teams at the global audio AI competition ‘CLIP (Cadenza Lyric Intelligibility Prediction) Challenge’ held in December last year. At ICASSP 2026, the company officially presented its achievements and technical details on the international academic stage.

The competition is known for its challenging tasks, which assess how accurately lyrics are conveyed in environments with mixed noise and music. Industry experts consider this a key indicator for evaluating the overall competitiveness of voice and auditory-based AI technologies.

CI Tech plans to sequentially apply this technology to its major business areas in the future. In particular, in the humanoid robotics sector, it is expected to serve as a core technology enabling accurate voice recognition and natural communication even in noisy industrial settings or complex public spaces.

The company will first introduce the technology to its affiliate, Ann Robotics, as well as other robotics firms, and then expand its application into the media business and AI healthcare service sectors.

CI Tech is actively pursuing stable growth in its existing businesses, as well as AI technology development and new AI healthcare ventures. The company explained that the ICASSP presentation and the victory at the global competition are evidence that its three core strategies have entered the full implementation stage.

A company representative stated, “We are doing our utmost to continuously strengthen our AI technology development capabilities so that we can achieve meaningful results not only for our affiliates but also in our own AI healthcare business in the near future.”

The representative added, “We believe that AI voice recognition technology will evolve beyond simple conversation recognition in noisy environments to the field of multi-speaker recognition, and we are focusing on developing related technologies as well.”

With its global competition victory and official presentation at ICASSP, CI Tech has proven its world-class AI technology in noisy environment voice recognition and is accelerating the expansion of its humanoid and healthcare businesses.

This content was produced with the assistance of AI translation services.