Earticle

In this paper, feature extraction and facial recognition are studied in order to resolve problems like high-dimension problem, small size samples and no-linear separable problem that exist in facial recognition technology. In the part of feature extraction we use a Discrete Cosine Transform (DCT) algorithm, to extract the input features in building a face recognition system. The RBF neural network, which represents brilliant performance on small training sets, non-linear separable and high-dimension pattern recognition problems in the recognition stage, is used for pattern classification. The proposed approach is validated with the ORL database. Experimental results demonstrate the effectiveness of this method in the performance of face recognition.

Query subtopic mining aims to find aspects to represent people’s potential intents for a query. Clustering query reformulations is the most common approach for subtopic mining these days. However, there are some challenges that the existing approaches have to face in finding both relevant and diverse subtopics, such as term mismatch and data sparseness. In this paper, a novel semantic representations for query subtopics is introduced, which including phrase embedding representation and query category distributional representation, to solve those problems mentioned above. Furthermore, we also combine multiple semantic representations into vector space model and compute a joint similarity for clustering query reformulations. To evaluate our theory an experiment is conducted on a public dataset offered by NTCIR subtopic mining project, the experimental results show that phrase embedding representation is the most effective representation while combining multiple semantics benefits short text clustering and improves the performance of query subtopic mining.

Due to advances in mobile networks and smartphone popularity in recent years, users can now connect to the internet anytime from anywhere and social searches are becoming popularly. Unlike conventional web searches, the social search provides optimal search results according to user preferences. In this paper, we propose a mobile social search method based on popularities and user preferences. The popularity is calculated by collecting the visiting records of users. The user preferences are generated by the actual visiting information among the search results. We process a skyline query to extract the meaningful information from the candidate objects with multiple features. The proposed method ranks social search results by combining user preferences and popularity with the skyline query processing mechanism. To show the superiority of the proposed method, we compare it with the existing method through performance evaluation.

CAPTCHA is an important technology to prevent auto-script attack. Currently most of the CAPTCHA systems are text based, which firstly distort, rotate different characters and then use some obfuscation, aiming to make the text difficult to be recognized by auto-script while still can be learnt easily by real users. However, such kind of CAPTCHA schema either too simple, which can be attacked easily by using optimal character recognition (OCR) or machine learning based technology, or it is too complex that even real users cannot tell it. By observing such contradiction between security and usability, we propose a novel object and text recognition based image CAPTCHA system called OTRCaptcha. In OTRCaptcha, some object images (each has a label) and their names are attached into a background image respectively. In order to pass this CAPTCHA, users have to identify all the object images, labels within those objects and their names. Besides, users also need to identify the semantic relationship between object and its name. Both the theoretical analysis and experiment result show that OTRCapthcha can provide both high security and strong usability.

Shape template matching is an important approach in object detection and recognition. In this paper, we propose a fast and novel object detection method, which represents edge map contours with salient points and retrieves the target object by using a backtracking method with two stages from coarse matching to fine matching. Our method has two main contributions. One is we propose the way to represent contour structure with salient points. The other is that the object detection method proposed can directly operate on the real images, which improves its practicability. According to the experimental results, the proposed approach improves the speed than previous works.

Snow is a common natural phenomenon that is frequently observed in winter. To simulate the effect of snow accumulation and melt, we generate exposure map from calculating occluding and shadow area using depth map. For simulating snow accumulation, we generate snow accumulated map using dispersed wind direction and occluder slope. We use light map calculating attenuation with distance on light source for simulating snow melt. To render snow in real-time, we use vertex displacement according to snow accumulation and melt. Also, we represent snow diffuse effect using perlin noise and snow boundary using stochastic sampling.

On the basis of concept, characteristics and related work of collaborative learning, this paper proposes an intelligent web collaborative learning system prototype based on multi-agent technology. At first, the paper introduces a collaborative learning framework which supporting group learning and illustrates the whole learning process. Secondly, the paper presents some key issues to implementation of the virtual learning environment, such as intelligent grouping, automatic question answering, knowledge management, and opinion mining. Preliminary learning practice shows that it is a practical and efficient way.

Chinese ancient art is extremely developed, and includes many kinds with different development levels, varieties. Due to the long time passed since the make time, many art products become blurred. Then, how to repair and processing them is a very important work. With the development of computer technology, the researchers conducted in-depth investigation on image recognition technology, and a lot of technologies have been developed, which can be used in the image processing. In this paper, a new combined recognition method for artistic image has been developed. Template matching method is used to match and correct the art image. This method can process the fuzzy art image into clear images. According to the practical verification, the validity of the method has been proven.