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

This work generalizes an existing framework of nonseparable oversampled lapped transforms (NSOLTs) to effectively represent complex-valued images.
The original NSOLTs are lattice-structure-based redundant transforms, which satisfy the linear-phase, compact-supported and real-valued property. The lattice structure is able to constitute a Parseval tight frame with rational redundancy and to generate a dictionary with directional atomic images.
In this study, a generalized structure of NSOLTs is proposed to cover complex-valued atomic images.

This paper extends the theory of the one-dimensional oversampled linear-phase perfect reconstruction filter banks (OLPPRFBs) developed by Gan et al. to multidimensional (MD) cases and proposes MD nonseparable oversampled lapped transforms (NSOLTs). NSOLTs allow us to achieve an overcomplete analysis-synthesis system with nonseparable, symmetric, realvalued, overlapping, and compact-supported filters. The proposed systems are based on lattice structures and the redundancy is flexibly controlled by the number of channels and downsampling ratio.

The flapping flight of bats can serve as an inspiration for flapping-wing air vehicles. Obtaining an understanding of bat flight requires detailed, occlusion-free kinematics data that can only be collected using large numbers of cameras. Here, we have explored the use of low-cost cameras with low frame rates that result in nonlinear, large-baseline motions in image space.

In microscopy, new super-resolution methods are emerging that produce three-dimensional images at resolutions ten times smaller than that provided by traditional light microscopy. Such technology is enabling the exploration of structure and function in living tissues such as bacterial biofilms that have mysterious interconnections and organization. Unfortunately, the standard tools used in the image analysis community to perform segmentation and other higher-level analyses cannot be applied naively to these data.

In this paper, motivated by the continuously increasing presence of intelligent agents in everyday life, we address the problem of expres-sive photorealistic audio-visual speech synthesis, with a strong focus on the visual modality. Emotion constitutes one of the main driving factors of social life and it is expressed mainly through facial expres-sions. Synthesis of a talking head capable of expressive audio-visual speech is challenging due to the data overhead that arises when con-sidering the vast number of emotions we would like the talking head to express.

Dynamic hand gesture recognition has attracted increasing interests because of its importance for human computer interaction. In this paper, we propose a new motion feature augmented recurrent neural network for skeleton-based dynamic hand gesture recognition. Finger motion features are extracted to describe finger movements and global motion features are utilized to represent the global movement of hand skeleton.

This paper presents an efficient convolutional neural net- work (CNN)-based multiple path search (MPS) algorithm to detect multiple spatial-temporal action tubes in videos. With the pass information and the accumulated scores generated by forward message passing, the new algorithm reuses these information to simultaneously find multiple paths in back- ward path tracing without repeating the search process. More- over, to rectify the potentially inaccurate bounding boxes, we also propose a video localization refinement scheme to further boost the detection accuracy.