Electrical Engineering

During recent years remarkable progress has been made in visual saliency modeling. Our interest is in video saliency. Since videos are fundamentally different from still images, they are viewed differently by human observers. For example, the time each video frame is observed is a fraction of a second, while a still image can be viewed leisurely. Therefore, video saliency estimation methods should differ substantially from image saliency methods. In this paper we propose a novel method for video saliency estimation, which is inspired by the way people watch videos. We explicitly model the continuity of the video by predicting the saliency map of a given frame, conditioned on the map from the previous frame. Furthermore, accuracy and computation speed are improved by restricting the salient locations to a carefully selected candidate set. We validate our method using two gaze-tracked video datasets and show we outperform the state-of-the-art.

Digital editing technology is highly popular as it enables to easily change photos and add to them artificial objects. Conversely, video editing is still challenging and mainly left to the professionals. Even basic video manipulations involve complicated software tools that are typically not adopted by the amateur user. In this paper we propose a system that allows an amateur user to performs a basic matchmove by adding an inlay to a video. Our system does not require any previous experience and relies on a simple user interaction. We allow adding 3D objects and volumetric textures to virtually any video. We demonstrate the method's applicability on a variety of videos downloaded from the web.

In a world where surveillance cameras are at every street corner, there is a growing need for synergy among cameras as well as the automation of the data analysis process. This paper deals with the problem of reidentification of objects in a set of multiple cameras inputs without any prior knowl-edge of the cameras distribution or coverage. The proposed approach is robust to change of scale, lighting conditions, noise and viewpoints among cameras, as well as object rota-tion and unpredictable trajectories. Both novel and tradi-tional features are extracted from the object. Light and noise invariance is achieved using textural features such as ori-ented gradients, color ratio and color saliency. A probabilis-tic framework is used incorporating the different features into a human probabilistic model. Experimental results show that textural features improve the reidentification rate and the robustness of the recognition process compared with other state-of-the-art algorithms.