Noise
Sound
Along with Cochl, Datumo took part in the process of builiding an environmental sound recognition AI system based on deep learning. Datumo collected audio data from different places, such as residential areas, restaurants, subway stations, and inside the subway cars, and from different events, such as subway announcements, car horns, and footsteps to expand the range of recognized audio types and improve accuracy.
Datumo provides high quality data for smarter AI. As part of Datumo's Data Sponsorship Program, Datumo cooperated with Cochl in building the following dataset.
Cochl delivers top-quality machine listening technology to solve issues and challenges around the world. Cochl provides technology through accessible cloud API and Edge SDKs, which adds hearing abilities to any device or application. Based on research, the startup aims to create machine listening abilites that mimic human auditory perception.
Cochl is a 2-year consecutive winner of IEEE DCASE ’17 ’18, the most prestigious competition in sound AI, and ranked first among 558 teams. Not only is Cochl the only Korean company to have been nominated as one of the Top 15 of Slush 100, it is also one of the top 4 AI startups in autonomous systems ranked by NVIDIA headquarters.
"Sounds from different places" will be shared as open datasets.
Sound of road traffic
ex. car horns, opening and closing car doors, arrival or subways, etc.
- 6,292 files
Sound from different places
ex. residential areas, restaurants, subway stations, inside the subway car, etc.
- 13,300 files
Other sounds
ex. background talking, talking over music, footsteps while walking and running, etc.
- 8,000 files
Collection of audio data was succeeded using the audio recording feature embedded in Datumo's mobile crowd-sourcing platform, Cash Mission.
Process of data preparation
- Development of POC to set standards for detailed inspections according to different categories.
- Construction of methods for highly accurate data collection and inspection process.
- Establishment of definitions of anticipated edge cases
- Collection and inspection of dataset
Data collection and labeling were completed using the mobile version of Datumo's crowd-sourcing platform, Cash Mission.
Audio files recorded according to different categories- sounds from various places and daily lives
Sample Data
Talking Over Music
- Subway Train Door
- Car start up
- Opening/closing car door
- Humming a song
Research and development of sound AI services
CC BY-SA
Reusers are allowed to distribute, remix, adapt, and build upon the material in any medium or format, even commercially, so long as attribution is given to the creator. If you remix, adapt, or build upon the material, you must license the modified material under identical terms.
https://creativecommons.org/licenses/by-sa/3.0/deed.en