(1) Deep Learning for Machine Intelligence

Single Depth Image Super-Resolution Using Convolutional Neural Networks

A Fast Deconvolution-Based Approach for Single Image Super-resolution with GPU Acceleration

Deep Feature Embedding Learning for Person Re-Identification Using Lifted Structured Loss

Automatic Segmentation and Cardiopathy Classification in Cardiac MRI Images Based on Fully Convolutional Neural Networks

TV-SVM: Support Vector Machine with Total Variational Regularization

(2) Stereoscopic 3D Video Processing

Perceptual stereoscopic video coding using depth-of-focus blur effects

Example-based video stereolization (EBVS) for 2D-to-3D conversion

Joint geodesic depth propagation and depth up-sampling

Visual comfort assessment and enhancement based on saliency and DIBR

Reliability-based discontinuity-preserving stereo matching

(3) Multimedia Annotation and Retrieval

Semantic annotation based on semi-supervised learning and constraint propagation

Kernel sparse representation-based classification using multi-objective optimization

Content-based summarization and retrieval for news and sports videos

Interactive image retrieval by active learning from relevance feedback

Videotext detection, segmentation, and recognition

(4) Image Super-resolution

Novel Bayesian deringing method using spatial-gradient-local-inhomogeniety prior

Face super-resolution based on convex optimization (l1-norm)

Parallelization of super-resolution reconstruction on GPU and multi-core platforms

Curvature-preserving super-resolution with gradient-consistency-anisotropic-regularization prior

Dictionary-based super-resolution with nonlocal total variation regularization

(5) Computational Photography

High dynamic range imaging by tone mapping and multi-exposure fusion

Specularity removal in color images

Interactive image segmentation by user interactions such as makers and touch

Power-constrained contrast enhancement for low power LCD displays

(6) Perceptual Video Coding and H.265

Perceptual image/video coding using human color perception

Perceptual rate distortion optimization using free energy principles and structural similarity

Blocking artifact reduction using sparse representation

Perceptual block merging for quad-tree based partitioning in H.265

Motion compensated prediction and interpolation in H.265

Projects

Xidian Media Lab has been supported by:

(1) Organization Department of the CPC Central Committee
(2) National Natural Science Foundation of China
(3) Ministry of Science and Technology
(4) China Postdoctoral Science Foundation
(5) Novatek Electronics
(6) Samsung SDS
(7) Huawei Technologies
(8) ZTE Corporation
(9) Institute of Information and Communication Technology Promotion (IITP)

We would like to acknowledge their financial support for our research.


©2015 Xidian Media Lab. All rights reserved.
Email: jushutaoxidian@163.com
Baidu
sogou