Speaker Identification Using AI Networks
SIAN is an Artificial Neural Network Model made to enroll and identify the person of interest uniquely through voice. It learns discriminative features for different voices through spectrogram inputs. It is used in various platforms for speaker identification and speech recognition, using voice as our new identity. This model was proposed in 2018 and it achieved an accuracy of 89% surpassing that of Oxford’s state-of-the-art model by 9% approximately. Currently, this is the best Voice identification model in the world and is capable of achieving 98% accuracy on the Vox Celeb Benchmark data set.