Moschopoulos, SpyridonFudos, IoannisKoritsoglou, KyriakosTatsis, GiorgosTzovaras, DimitriosPelechano, NuriaLiarokapis, FotisRohmer, DamienAsadipour, Ali2023-10-022023-10-022023978-3-03868-233-2https://doi.org/10.2312/imet.20231250https://diglib.eg.org:443/handle/10.2312/imet20231250This paper reports on the development of a real-time voice interface for navigation purposes of electric wheelchairs. To this end, we employ a convolutional neural network trained and fine-tuned using a small dataset that consists of Greek commands. Furthermore, the study explores a highly quantized version of the network to achieve computational efficiency while maintaining high accuracy on an edge device. The experimental results confirm the effectiveness of the model in accurately detecting keywords in real time with minimal misclassifications.Attribution 4.0 International LicenseCCS Concepts: Computing methodologies -> Speech recognition; Supervised learning by classification; Transfer learning; Computer systems organization -> Real-time system architecture; Hardware -> Hardware acceleratorsComputing methodologiesSpeech recognitionSupervised learning by classificationTransfer learningComputer systems organizationRealtime system architectureHardwareHardware accelerators10.2312/imet.2023125019-224 pages