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

In this paper, we propose a new characteristic measure for relative people density and motion dynamics for the purpose of long-term crowd monitoring. While many related works focus on direct people counting and absolute density estimation, we will show that relative densities provide reliable information on crowd behaviour. Furthermore, we will discuss the derivation of a so-called Congestion Level of local areas in the crowd, which takes the current dynamics and density within a certain image region into account.

Numerous trackers have been proposed in recent years with considerable success. But few trackers can cope with all scenarios without failures. It is very difficult to design a tracker robust enough to keep off tracking failure. As failure is inevitable, we propose a framework to correct tracker, verify failure, predict object position and re-detect object. The original model of the first frame is used to correct the tracker. Then, the confidence of tracking is used to verify tracking failure.

Measuring the Heart Rate (HR) plays an important role in the description of human physiological and psychological state, due to its relationship with cognitive/emotional factors such as attention effort, stress or arousal. For this reason, remote methodologies for HR measurements have recently been investigated to find a reliable and cost-effective methodology. Our work aims at the following:
• Development of a novel technique for remote HR estimation
• Comparison of the proposed method with the state of the art on a common dataset

Facial Parts Detection (FPD) approach in conjunction with Vector Quantization (VQ) algorithm are proposed for face recognition. Detecting facial parts, which are nose, both eyes, and mouth, and choosing appropriate dimensions for each part, are done in the preprocessing phase. In the feature extraction phase, four groups for each person, one group for each detected part, are constructed for dimensionality reduction and feature discrimination by considering all parts of all training poses. For further data compression, VQ algorithm is applied to each of the four groups.

Authentication by facial recognition is actually one of the solutions to reinforce the security level
of information systems. However, face recognition systems are proven to be vulnerable to spoofing
attack. In fact, an attacker can bypass the authentification process easily by presenting in front of the
camera a copy version of a legitimate user’s face.
To make your face as your password, it is of vital importance to identify and reject the falsified faces

A Wiener filtering scheme in graph Fourier domain is proposed for
improving image denoising performance achieved by various spectral
graph based denoising methods. The proposed Wiener filter is
estimated by using graph Fourier coefficients of the noisy image after
they are processed for denoising, to further improve the already
achieved denoising accuracy as a post-processing step. It can be estimated
from and applied to the entire image, or can be used patchwise
in a locally adaptive manner. Our results indicate that the proposed

Saliency measures are a popular way to predict visual attention. However, saliency is normally tested on sets of single resolution images that are unlike what the human vision system sees. We propose a new saliency measure based on convolving images with 2D gamma kernels which function as a comparison between a center and a surrounding neighborhood. The two parameters in the gamma kernel provide an ideal way to change the size of both the center and the surrounding neighborhood, which makes finding saliency at different scales simple and fast.

Action recognition via WiFi has caught intense attention recently because of its ubiquity, low cost, and privacy- preserving. Observing Channel State Information (CSI, a fine-grained information computed from the received WiFi signal) resemblance to texture, we transform the received CSI into images, extract features with vision-based methods and train SVM classifiers for action recognition. Our experiments show that regarding CSI as images achieves an accuracy above 85%. Our contributions include: