Goodness of Pronunciation Algorithm in the Speech Analysis and Assessment for Detecting Errors in Acoustic Phonetics: An exploratory review

The principal objective of this study is to evaluate the GOP algorithm proficiency to carry out a portable device based on evolutionary algorithm classifier integrated in an embedded system for children mispronunciation diagnosis in phonoaudiology therapy. GOP applications are widely worked in foreign language learning, however, the background around the medical trends shows the usage extensibility in speech disorders diagnosis and therapy. This review is an initial part of our future project where the GOP algorithm will be evaluated by its implementation in embedded hardware systems like artificial intelligence Single Board Computers (AI-SBC) based on a Tensor Processing Unit (TPU) or a Neural Processing Unit (NPU). Furthermore, a new proposal to integrate GOP score machine with evolutionary algorithms, for example, Learning Classifier System approach (LCS), will be designed in an ubiquitous model for phonoaudiology usage in children speech impairments detection.