PYNQ-based intelligent music player
Li He, Xie Kanghui, Wang Chengcheng, Tang Yongming(Mentor)
This work has designed an intelligent music player that can play appropriate music according to your facial expression or you voice. The system consists of speech recognition module, facial expression recognition module, gesture contral module and sound playback module. And our functions are as follows:
1). recognize the face expression and its probability;
2). recognize user's voice, which will be converted into text and output by screen;
3). play music based on facial expression recognition;
4). play music based on voice recognition;
5). establish human interaction system based on gesture recognition;
The work has a wide application prospect in the future. For example, it can connect to the home network as a music player and be used as a member of the Internet of things. At the same time, if the users are far away from the device, we can also control it using gesture. Since the system can be independently networked, it is also very convenient for subsequent development and application.
The innovation points of this work are:
1)We make full use of the integrated development advantages of pynq-z2 for image manipulation and speech recognition, so that the accurate is highly improved;
2)We use Basys3 for gesture contral module. With the help of logical design, on the one hand, we successfully programmed UART and I2C bus protocols which are based on Verilog; on the other hand, we finished the high-speed processing of gesture data. So that FPGA's flexibility and high degree of freedom in timing design are highly utilized.
3)We realize the offline operation of the production. It can exist as an independent product with greater practicality and presentation.
System Construction & Function Description
1)overview flow chart
gesture contral module
speech recognition module
facial expression recognition module
System Interface diagram
We select some representative pictures from the syetem: