Big Data
Sound Collection
- Reading Recordings~Collection
- Collect sound data by reading news reports, papers and other manuscripts.
- Original Production ~ Recordings:
- Create the most suitable plot manuscript for recordings.
- Dialogue script production:
- Setting the scene to fit the purpose of the utilization and creating a natural script for the conversation.
- Recording method:
- Client-supplied equipment, systems, etc.
- Corresponding Languages:
- Hindi, Japanese, English, Chinese, Korean, etc.
Handwriting Collection
- Corresponding Languages:
- Japanese, Chinese (Simplified, Traditional), English, Hindi, etc.
- Operating Environment:
- Tablet. smart phone, etc. (online, offline)
Image or Photo Collection
- Investigation and Collection:
- Photo collection, shopping collection, transportation advertisement related collection, etc.
- Specific Collections (Large Projects):
- Collection Quantity:
- Number of People to Collect
- Collection Requirements:
- ① Light, Angle and Background
② Time, props, posture and expression, etc.
- Delivery:
- Client's Specified
- Operating Environment:
- Client-specified devices or systems such as smart phones or tablets
Labelling
- Sound Labelling:
- Labelling of sound waveforms and determination of sound validity through dedicated applications
- Handwritten Labelling:
- Write by hand and add attributes to the labels to determine whether the handwritten content is valid.
- Portrait or Photo Labelling:
- Facial Photos:
- 117 dots, 216 dots, cat face, face recognition sign
- Posture Photo:
- 17 dots, 26 dots, frame or shelter (with or without)
Photo survey (shopping, electrical goods, magazines, books, shop billboards, etc.) → more frames on display
- Operating Environment:
- Client-provided, client-specific applications