RNN-based Dialogue Navigation System for Visually Impaired
Ching-Han Chen, Ming-Fang Shiu · 2020
Helping the visually impaired to walk and guiding him to the destination is a challenging task. The difficulty is to use natural language as the only communication method and assist the visually impaired to see the road. We have developed a conversational navigation system that integrates multiple rounds of dialogue to confirm the destination for the visually impaired. Besides, we also use image interpretation technology based on the RNN neural network to describe the scene in front of the user. Finally, in terms of hardware, we use a low-cost, low- power embedded hardware that integrates cameras, Wi-Fi, and microphones to implement this application.