In order to build an efficient, sample and convenient system of face recognition based on the video streaming, a mixed mode of software development is proposed in this paper. In requirement analysis of the face recognition system, efficient code, sample development and ease of use are designed in this system. To obtain better recognition efficiency, Dynamic Link Library (DLL) is used to program the part of face recognition. The whole system is based on Microsoft .NET platform which provides a simple and rapid development. The mixed of Browser/Server (B/S) and Client/Server (C/S) modes makes the system more flexible, at the same time more convenient for user to use. The whole system is divided into three parts: server, client and browser. And the server-side is storing data through a database, managing different clients and providing data to the browsers. The client-side mainly monitors and collets data. The browser-side provides users with querying information which they want. And the proposed system architecture can satisfy the application requirements for the customer.
Flame Detection based on video is an important method for fire prevention. To save flame detection time, the moving objects were segmented from images in a video by using GMM(Gaussian Mixture Model) firstly. Then moving objects were determined as flame candidate or not by their color feature and area changes. The experiments show that the method could detect flame in video effectively with a low false positive rate.
Fire detection based on images is an effective method for fire prevention, especially in big room or badly environment. It is important to extract the features of a flame image. According to the idea of visual saliency in computer vision, saliency model of brightness, color and flame texture are proposed here. The saliency of flame brightness is indicated by the V component in HSV color space. The saliency of flame color is expressed by the relation of R and B in the RGB color space. The saliency of flame texture is described by the distance between the feature vectors which are the combination of features with Local Binary Pattern. Experimental results show the saliency model is effective for flame feature extraction.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.