SPS Feed

Visual attention is an important mechanism in the human visual system (HVS) and there have been numerous saliency detection algorithms designed for 2D images/video recently. However, the research for fixation detection of stereoscopic video is still limited and challenging due to the complicated depth and motion information. 

This paper presents a new method of secret three-dimensional object sharing (S3DOS), which allows sharing of three-dimensional (3-D) objects, while preserving its file format by selectively encrypting a 3-D object in order to sufficiently protect the visual nature of the content. 

This paper addresses the problem of encoding the video generated by the screen of an airplane cockpit. As other computer screens, cockpit screens consist of computer-generated graphics often atop a natural background. Existing screen content coding schemes fail notably in preserving the readability of textual information at the low bitrates required in avionic applications. 

Image classification is an essential and challenging task in computer vision. Despite its prevalence, the combination of the deep convolutional neural network (DCNN) and the Fisher vector (FV) encoding method has limited performance since the class-irrelevant background used in the traditional FV encoding may result in less discriminative image features.

Dictionary learning for sparse representations is generally conducted in two alternating steps-sparse coding and dictionary updating. In this paper, a new approach to solve the sparse coding step is proposed. Because this step involves an ℓ0 -norm, most, if not all, existing solutions only provide a local or approximate solution. Instead, a real ℓ0 optimization is considered for the sparse coding problem providing a global solution. 

Short duration text-independent speaker verification remains a hot research topic in recent years, and deep neural network based embeddings have shown impressive results in such conditions. Good speaker embeddings require the property of both small intra-class variation and large inter-class difference, which is critical for the ability of discrimination and generalization.

Representation learning is the foundation of machine reading comprehension and inference. In state-of-the-art models, character-level representations have been broadly adopted to alleviate the problem of effectively representing rare or complex words. However, character itself is not a natural minimal linguistic unit for representation or word embedding composing due to ignoring the linguistic coherence of consecutive characters inside word.

In this letter, we propose a heuristic method to address sensor bias estimation to improve track-to-track association accuracy. A novel multi-parameter cost function is derived from rigid transformation function and it is minimized by the covariance matrix adaptation evolution strategies algorithm.

We study the problem of distributed filtering for state space models over networks, which aims to collaboratively estimate the states by a network of nodes. Most of existing works on this problem assume that both process and measurement noises are Gaussian and their covariances are known in advance. In some cases, this assumption breaks down and no longer holds.