Vision and Learning (VL) Lab.

We are interested in various problems in computer vision and machine learning. Particularly, our recent studies focus on studying 1) generative models toward image synthesis, image editing, face modeling and recognition, 2) deep neural networks for object detection and classification, 3) 3D modeling and reconstruction, inverse rendering, 4) retrieval, recommendation, and hashing algorithm and 5) medical image processing.

Current Projects

https://sites.google.com/site/cvgyonsei/home/Thumbnail_ICML.png

Resembled GANs

Simultaneously generating two domain images with similar attributes.

Editing the attributes of existing images or generating images with desired attributes simultaneously through single training procedure

Unsupervised object localization through self-supervised approach

Proposing Attention-based Dropout Layer (ADL), which utilizes the self-attention mechanism to process the feature maps of the model

Proposing a new evaluation protocol for WSOL

3D Reconstruction

Fully automated depth correction from Kinect V2 depth camera with ResNet.