-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Safe Noise Canceling is an application that detects dangerous sounds around you and alerts you when using sound devices such as headphones.
Currently, it is only available on Android platforms.
This project is of a research and does not guarantee stability in actual use. Therefore, all risks arising from actual use are the responsibility of the user.
According to market research firm Gallup Korea, Korea's smartphone usage rate surpassed 70% in February 2013, 80% in July 2014, and 90% in the second half of 2016, from 53% in 2017 to 2020. This increase in smartphone use has also resulted in an increase in the use of portable sound devices such as earphones. In addition, wireless earphones using Bluetooth technology began to appear in 2016, making it more convenient to use.
The increase in the supply of these devices has given users a lot of convenience, but it has also begun to have a great impact on traffic safety. According to research data released by the Korea Highway Traffic Authority in 2015, 27.9% of respondents said they had almost experienced an accident due to listening to sound devices while walking, and 1.6% said they had experienced the accident. According to Heinrich's law, the researchers who conducted this study said that 29 minor events occurred and 300 dangerous situations occurred when one major accident occurred, and this study showed a similar trend.
Recently, with the emergence of earphones with technology such as noise cancelling that completely blocks surrounding sounds, the risk of using sound devices while walking is higher than in 2015 when the study was conducted.
After seeing this increase in risk, our team decided to devise a system that listens to and recognizes dangerous sounds around us instead of users and notifies them.
- Recognizes dangerous sounds heard around and alerts the user.
- Listen to the surrounding sound on the smartphone and analyze the sound through the TensorFlow Lite (YAMNet).
- Control the notification frequency according to different situations by changing the recognition rate according to the user's current behavior
- Microphone calibration to support various devices.
This project is an Undergraduate Graduation Project of the School of Computing, Gachon University in South Korea.
Advisor : Jaehyuk Choi
Team Member : Nahyeong Kim, Dajeong Park, Minjae Seon