This paper presents a technique to recover geometry from time-lapse sequences of outdoor scenes. We build upon photometric stereo techniques to recover approximate shadowing, shading and normal components allowing us to alter the material and normals of the scene. Previous work in analyzing such images has faced two fundamental difficulties: 1. the illumination in outdoor images consists of tim...

In this paper we present a novel approach for interactive rendering of virtual views from real image sequences. Combining the concepts of light fields, depth-compensated image warping and view dependent texture mapping, this plenoptic modeling approach can handle large and complex scenes. A portable, handheld multi-camera system has been developed that allows to record multiple image streams by...

Mosaic images provide an eecient representation of image sequences and simplify scene exploration and analysis. However, the application of conventional methods to generate mosaics of scenes with moving objects causes integration errors and a loss of dynamic information. In this paper a method to compute mosaics of dynamic scenes is presented addressing the above mentioned problems. Moving pixe...

A wearable eye tracker was used to record photographers’ eye movements while they took digital photographs of person, sculpture, and interior scenes. Eye movement sequences were also recorded as the participants selected and cropped their images on a computer. Preliminary analysis revealed that during image capture people spend approximately the same amount of time looking at the camera regardl...

Object classification in far-field video sequences is a challenging problem because of low resolution imagery and projective image distortion. We approach the problem by identifying discriminative object features for classifying vehicles and pedestrians in far-field video captured by a static, uncalibrated camera. Using these features, we design a scene-invariant classification system that is t...

Participants viewed slides depicting ordinary routines (e.g., going grocery shopping) and later received a recognition test. In Experiment 1, there was higher recognition confidence to high-schema-relevant than to low-schema-relevant items. In Experiment 2, participants viewed slide sequences that sometimes contained a cause (e.g., woman taking orange from bottom of pile) but not an effect scen...

This paper presents some techniques for automatically deriving realistic 2-D scenes and 3-D geometric models from video sequences. These techniques can be used to build environments and 3-D models for virtual reality application based on recreating a true scene, i.e., tele-realityapplications. The fundamental technique used in this paper is imagemosaicing, i.e., the automatic alignment of multi...

In this article we consider the sequences of sample and population covariance operators for a sequence of arrays of Hilbertian random elements. Then under the assumptions that sequences of the covariance operators norm are uniformly bounded and the sequences of the principal component scores are uniformly sumable, we prove that the convergence of the sequences of covariance operators would impl...

To combat with lagging rate control (RC) parameter setting in encoding video sequences with discontinuous scenes, a novel RC algorithm at group of picture (GOP) level is proposed for high efficiency video coding (HEVC) in this paper. Firstly, the variation of GOP is analyzed and used to detect the discontinuous scene. Then, a new bit allocation algorithm is presented by constructing the correla...

We present an efficient approach that merges the virtual objects into video sequences taken by a freely moving camera in a realistic manner. The composition is visually and geometrically consistent through three main steps. First, a robust camera tracking algorithm based on key frames is proposed, which precisely recovers the focal length with a novel multi-frame strategy. Next, the concerned 3...