Speech Sentence Recognition Guidance System

Speech Sentence Recognition Guidance System

A recently research, Speech recogntion AI model was developed to integration with system such as IoT (Internet of Things), management system and devices. so this paper indroduced to Speech sentence recognition AI guidance system. The speech recognition basic model was classifier to simple word. but as researching speech recognition topic, the speech recognition model had powerful performance what model can classifier to difficult sentence in speck. these powerful speech recognition model was provided to convinience while converging with various devices that we use in real life. In particular, smartphones, which are computers carried by modern people, are designed to control functions that can be used in the system when emergency situations and daily life by commanding by voice as well as voice recognition. Our system uses these speech recognition system and camera detection capabilities to propose an unmanned parking management system. this is expected to save labor costs and bring convenience along with strong performance by making it available 24 hours a day. the speech recognition system used in this paper introduces the speech signal processing technologies of Spectrogram, Mel spectrogram and MFCC. These speech signal processing techniques play an important role in the model as a technology for extracting information contained in speech, such as the accent, pitch and parttern of the speaker's pronunciation. Moreover, YOLO(Your Only Look Once), which is widely used in the field of object recognition, not only provides strong object recognition, but also has accuracy while consuming less resources on the device. We expect to bring many advantages by builing a system that can integrate and manage these two model. These two systems require a framework for applications. therefore, we built applications using the QT framework and designed it to be used in devices and PC. the QT framework is a strong cross-platform framework and can be used in various operating system. Therefore, we constructed the system by fusing these two models with the application and conducted research with various function required accordingly.