AI-based media processing research
We conduct research on the visual intelligence of computers,
training them to extract, understand, analyze various information
from images and videos to identify and solve problems.
Face Swap
Transcending the limits of content creation. We lead the innovation in the content creation environment with the generative adversarial neural network (GAN)-based media processing technology.
Face Swap Process
Face Generation
The person in your imagination becomes a reality. Using a huge dataset of images of various people, AI can generate photorealistic images of people that do not exist in real life.
Face Generation Process
Image Inpainting
Magically remove any unwanted objects! High-quality results can be achieved in a short time by automatically removing undesired objects from high-resolution images and videos and restoring the images and videos seamlessly.
Image inpainting Process
Video Abstraction
Fast and smart video clip creation. Face and object detection, sentiment analysis, and audio signal analysis are applied to various video content, such as broadcasts, live commerce, and sports broadcasts, to automatically summarize videos.
Video abstraction Process
Style Transfer
From morning to night, summer to winter, real life to animation, AI technology makes it possible to build an environment without the constraints of location by converting the time of day, season, and features in the original video and image.
Style Transfer Process
Anomaly Detection
Preemptively resolving potential issues before they become real issues. With the vision AI-based anomaly detection technology, defects in the production process and labeling errors can be detected and preemptively resolved.
Anomaly Detection Process
Vision AI has a wide variety of uses.
Face Swap,
Face Aging/De-Aging
The GAN-based face synthesis technology increases the liberty of content creation, reduces costs, while facilitating viewer immersion and delivering stronger emotions. The technology that modifies or removes a specific person appearing in an existing video enables a quick and efficient response to a social controversy.
Virtual Influencer
Create your own virtual influencer with a realistic personality that has never existed before. Using a face generation technology, you can turn a person in your imagination into a virtual influencer of any race, gender, and style.
Object Eraser
When editing to remove trademarks or signboards in a video, you can reduce both the production time and cost by using an automated AI model. Objects such as logos and signboards that should not be exposed in commercial broadcasting are automatically detected and removed, and the removed parts are reconstructed to fit into the surrounding environment to create a high-quality video.
Media content preview and automatic creation of key video clips
Create video clips that can satisfy your clients based on image, audio and sentiment analysis. The key video clips are automatically summarized and created by analyzing the positivity of the person's speech in the video, object sections, object identification, banner clicks, and chat data.
Smart factory
Producing smart products with a smart factory The AI defect detection model incorporated in the smart factory detects and visualizes the locations of defects to automatically classify defective products and improve the defect rate.
Research
papers
Nakyung Lee et al., 2021, "A Body Shape Change Visual Tracking Service Using Deep Learning-Based Human Part Segmentation," 2021 Fall Conference of the Korea Society of IT Services
DownloadJungsu Hwang et al., 2021, "Live-Commerce Video Summarization Using Analysis of Multimodal Information Based on Deep Learning," the Journal of Digital Contents Society
DownloadPatent
applications
Would you like
to inquire now?
Please submit your inquiry,
and we will respond as soon as possible.
No attachment found
* You can send up to 10 MB in attachments. Supported file formats: zip, pdf, hwp, ppt, pptx, doc, docx, xls, xlsx, jpg, jpeg, png, gif
Contact Us
No attachment found
* You can send up to 10 MB in attachments. Supported file formats: zip, pdf, hwp, ppt, pptx, doc, docx, xls, xlsx, jpg, jpeg, png, gif