최광남 연구실

Currently, deep learning is a technology that is in the spotlight in various fields. We use deep learning effective in processing images such as object detection and image segmentation to improve research results.

GAN is a deep learning technology that generates data and is widely used in many fields, such as medicine and fashion, where data is scarce. We are researching with the aim of creating pedestrian images using GAN.

Centernet is the most recently proposed network with high accuracy and high speed. We use centernet model and train the kitti 3D dataset to detect 3D objects.

We utilized deep learning to perform landmark detection, which is widely used in fashion. This experiment was performed by adding keypoints to the efficientnet, and used the fashion 2 dataset.

Image segmentation is a principal research direction in image processing and is widely used in image analysis, medical imaging and computer vision. Active contours are deformable models used in medical image analysis.