Solution outline
Deep network models have been trained on the custom datasets to detect anblur vehicle license plates andhuman face in the wild.
This solution can be effectively usedfor security and hiding the identity of people in recorded videos.
:Vehicle license plate and face blurring
AI detects and blurs the vehicle license plates and face in the wild
to hide person identity from the videos
Solution outline
The neural network model has been trained on various accident cases for the purpose of
finding and detecting t he accident instances from videos.
This solution detect and mark the accident instances from the full length raw videos recorded from the dash cams.
:Accident detection AI detects
if there are any accident instances
from the input raw video recorded from the dash cam
Solution outline
Using data collected through various sensors, images of various objects are recognized and judged,
and names and characteristics are determined accordingly.
: Item recognition using 1-stage detector (YOLO).
AI detects if there are dangerous items
in the collected images.
Solution outline
It is possible to measure the amount of liquid by calculating the volume of an object
with a hidden part by using AI to calculate the invisible part.
In addition, by restoring the body in 3D, it is possible to reconstruct lost body parts congenitally or acquired.
: Background image object using point cloud,
image pre-processing process to separate background and object -> 3D object modeling
Solution outline
In the ever-changing world of fashion, you can find fashion products
through images and recognize/classify product categories.
In addition, AI technology helps
in analyzing social media data to identify trends of influencers and stars.
: Detect current user's clothes using AI and analyze
the color and combination of clothes in the latest trend
Solution outline
Recognize/determine objects such as movement of various objects and obstacles by using data collected through various sensors.
It helps the driver to drive more safely along the driving route.
: I: convert image space to hough space
Solution outline
In order to reduce greenhouse gas and carbon emissions from cement,
construction companies can search and connect with companies that operate in an eco-friendly way.
For cement manufacturers, we have established a system
that can calculate carbon emissions and predict the amount of raw/subsidiary materials.
: After using the oversampling technique (SMOTE, ADASYN)
to solve the data imbalance problem Utilization of Gaussian noise data augmentation
to reduce data size gap between data to be compared