A new generation of Google may not be far away, the future of search will rely on computer vision – w-inds.

A new generation of Google may not be far away, the future of the search will rely on computer vision – Sohu technology in the streets of San Francisco, someone will use a Alice Who (Alice) APP. You just put the mobile phone camera at a building, or a restaurant, "Alice Who" will tell you, what is this place, what’s the name of the restaurant called "Alice Who"; also dianping.com Yelp, help you connect to the American version of the FourSquare, so you can understand the detailed information and evaluation this restaurant. With Alice Who, local search becomes very simple, like a local friend in the side, where you look, the local friends can tell you what to see. This may be the future search experience! Seemingly simple applications, behind the complex technology. For outdoor environment in the building, the identification of the store, has been the difficulty of computer vision. How to make the computer to see clearly, see accurate? Even if the outdoor light changes, even if there are many interference factors in the environment? With these questions, the Silicon Valley Alice Who developers to connect the spy, Israel innovation company Fringefy co-founder Assif Ziv, please he introduced Fringefy technology in the field of computer vision research, and Fringefy’s vision and direction. A computer vision computer vision technology (Computer Vision) in recent years, we talk about many topics. Unmanned vehicles, unmanned aerial vehicles and other products gradually into the public view, how to make the machine like people, have the ability to see and identify the surrounding environment, the key to the success of these products. With the development of computer vision technology, different methods have been developed to help the machine "see" objects. Most of the technology is based on the same foundation, which is the point cloud (Point Cloud). Through the 3D scanning objects, to build a point cloud, and then use a point cloud training machine algorithm, so that the machine can identify the object. (Point Cloud schematic) but this standard technique in indoor application, object recognition, face is very awesome, but not good at identifying outdoor building. In the outdoors, with the change of the light, not only the color and intensity of the building will change, the shadow will change; in addition, the pedestrian around the building, parked vehicles, will also affect the identification of the machine. (outdoor environment and object is affected by many factors, images from asl.ethz.ch) for example, we in the morning 10 point shooting coffee images, the images again shot 4 points and the afternoon, for the machine recognition, may be two completely different scenes. Moreover, the larger the size of the building, people usually can not take the picture. It also poses a challenge to machine recognition. Computer vision technology and Fringefy a, mainly focus on addressing these challenges. Fringe.相关的主题文章: