Sorry, you need to enable JavaScript to visit this website.

Retinex theory deals with compensation for illumination effects in images, which is usually an ill-posed problem. The existence of noises may severely challenge the performance of Retinex algorithms. Therefore, the main aim of this paper is to present a general variational Retinex model to effectively and robustly restore images corrupted by both noises and intensity inhomogeneities. Our strategy is to simultaneously recover the noise-free image and decompose it into reflectance and illumination component.

The deep convolutional neural network(CNN) has significantly raised the performance of image classification and face

Dual-fisheye lens cameras are becoming popular for 360-degree video capture, especially for User-generated content (UGC), since they are affordable and portable. Images generated by the dual-fisheye cameras have limited overlap and hence require non-conventional stitching techniques to produce high-quality 360x180-degree panoramas. This paper introduces a novel method to align these images using interpolation grids based on rigid moving least squares. Furthermore, jitter is the critical issue arising when one applies the image-based stitching algorithms to video.

Although for a still image the 2-D DFT and the 2-D DCT have similar properties to each other, for a moving-image sequence the 3-D DFT gets an advantage of representing the sequence more compactly over the 3-D DCT. Through the mathematical analysis of the 3-D DFT and the 3-D DCT based on a simple signal model of a moving-image sequence, this paper shows that the even symmetrization employed implicitly by the 3-D DCT may cause deterioration of representation efficiency and hence the 3-D DFT can achieve better representation efficiency than the 3-D DCT.

Active contours based on level sets are popular segmentation algorithms but their local optimization approach makes their results to depend on initialization, especially for edge-based formulations. In this paper we present a novel energy minimization method based on directed graph optimization that minimizes the same type of active contour energy function without the need of an initialization.

We look at the problem of developing a compact and accurate model for gesture recognition from videos in a deep-learning framework. Towards this we propose a joint 3DCNN-LSTM model that is end-to-end trainable and is shown to be better suited to capture the dynamic information in actions. The solution achieves close to state-of-the-art accuracy on the ChaLearn dataset, with only half the model size. We also explore ways to derive a much more compact representation in a knowledge distillation framework followed by model compression.

Circle detection is fundamental in both object detection and high accuracy localization in visual control systems. We propose a novel method for circle detection by analysing and refining arc-support line segments. The key idea is to use line segment detector to extract the arc-support line segments which are likely to make up the circle, instead of all line segments. Each couple of line segments is analyzed to form a valid pair and followed by generating initial circle set.