The objective of this work is to recognize object categories (such as animals and vehicles) in paintings, whilst learning these categories from natural images. This is a challenging problem given the substantial differences between paintings and natural images, and variations in depiction of objects in paintings [5] – see figure 1. Contributions. (i) We show that object category classifiers lea...

A model for tracking articulated objects is proposed using a novel 2D revolute-prismatic joint. An extension of the RANSAC and MLESAC algorithms incorporating feature weights is used to perform robust tracking. The models are suitable for tracking certain human body structures. Limbs are modelled as constrained planar patches. A patch can rotate about a joint point that is displaced relative to...

In this paper, we proposed a new feature based image mosaic algorithm. The improved RANSAC homography algorithm based on the modified media flow filter, to detect wrong matches for improving the stability of the normal RANSAC homography algorithm. The method improved the local registration between neighboring images. Experiments and Statistical Analysis show that our mosaic method is robust. Ke...

Obstacle avoidance is a major requirement for any technological aid aimed at helping partially sighted (TAPS) people to navigate safely. In this paper, a stereo vision-based algorithm (Ground Plane Obstacle Detection) is extended to detect small obstacles for TAPS using RANSAC dynamic recalibration and Kalman Filtering. Obstacle detection and false alarm are investigated probabilistically. Furt...

In this paper, we propose a new method to compensate for lens distortions in image stitching. Lens distortions that arise from the nonlinearity of a lens are the main cause for mismatches in stitching images. We estimate the distortion factors for each image using the Division Model and linearize the projected relationships between matching distorted feature points. Because our method works at ...

This paper addresses the 3D tracking of pose and animation of the human face in monocular image sequences using deformable 3D models. For each frame, the proposed adaptation is split into two consecutive stages: global and local. In the first stage, the 3D pose of the face is recovered using a RANSAC-based technique involving both the consensus measure and the consistency with a statistical mod...

In this paper we will consider a combination of the RANSAC algorithm and the Hough transform for fast model estimation under the presence of outliers. The model will be computed by sampling a smaller than minimal subset, followed by a voting process of the remaining data. In order to use the combined method for this purpose, an adequate parameterization of the model in the Hough space is requir...

Usual Structure-from-Motion (SfM) techniques require at least trifocal overlaps to calibrate cameras and reconstruct a scene. We consider here scenarios of reduced image sets with little overlap, possibly as low as two images at most seeing the same part of the scene. We propose a new method, based on line coplanarity hypotheses, for estimating the relative scale of two independent bifocal cali...

We present a fast, robust method for registering successive laser rangefinder scans. Correspondences between the current scan and previous scans are determined. Gaussian uncertainties of the correspondences are generated from the data, and are used to fuse the data together into a unified egomotion estimate using a Kalman process. Robustness is increased by using a RANSAC variant to avoid inval...

Terrestrial laser scanning (TLS) is a promising technique for plot-wise acquisition of geometric attributes of forests. However, there still exists a need for TLS applications in mountain forests where tree stems’ growing directions are not vertical. This paper presents a novel method to model tree stems precisely in an alpine landslide-affected forest using TLS. Tree stems are automatically de...