Data Service
It has been important to construct high-quality mass datasets as AI transformation has expanded
across various
industrial areas. INFINIQ has accelerated such a transformation based on the DataStudio
framework and provides
the dataset construction service best fit for the whole data life cycle
across data
design/collection/cleansing/labeling and dataset quality control.
Data Collection
Construct the source data available for AI recognition use.
- Service Features
-
High-precision data collection such as multi-modal
data acquisition according to various environments,
scenarios, and sensor fusion
- How to Collect
-
Support of driving logging, crawling, production, video
recording, studio running, and 2D+3D sensor fusion
calibration
Data Cleansing
Vast source-data characteristics are sorted to prepare effective target data.
- Service Features
-
Characteristics are auto-sorted with techs of data
virtualization and computer vision used.
- How to Cleanse
- Data profiling, data curation, automatic deidentification
Data Annotation
GT is created by applying labeling methods best fit to AI algorithm learning.
- Service Features
-
A labeling management solution into which the
knowledge of vision AI and labeling operation is
integrated Productivity and quality level maximized
with MyCrowd
- How to Annotate
-
Meta tagging, BBox, Polygon, Semantic
Segmentation, Polyline,Cuboid, Point, Skeleton,
LiDar(3D PC)...
Data Inspection
Quality Control to secure quality metrics and quality goals of a dataset.
- Service Features
-
Inspection with computer vision technologies and
by data quality experts
- How to Inspect
-
Inspection of compatibilities in meaning and syntax according to Dataset Quality Standards
Systematic quality control about typical error types
: Object missing, fit, class mismatch