Geometric fusion of multiple sets of overlapping surface measurements is an important problem for complete 3D object or environment modelling. Fusion based on a discrete implicit surface representation enables fast reconstruction for complex object modelling. However, surfaces are represented at a single resolution resulting in impractical storage costs for accurate reconstruction of large obje...

In the last few years the research in 3–D object recognition has focused more and more on active approaches. In contrast to the passive approaches of the past decades where a decision is based on one image, active techniques use more than one image from different viewpoints for the classification and localization of an object. In this context several tasks have to be solved. First, how to choos...

Hyperspectral imagery contains a wealth of spectral and spatial information that can improve target detection and recognition performance. Typically, spectral information is inferred pixel-based, while spatial information related to texture, context and geometry are deduced on a per-object basis. Existing feature extraction methods cannot fully utilize both the spectral and spatial information....

The Fusion and Syntropy Object-Oriented methods reflect some of the most rigorous object-oriented modeling concepts and experiences currently available. In this paper we identify similarities and differences between the two methods, and discuss how the best concepts from each of these methods can be combined to obtain clearer, more precise analysis models, specifications, and designs, and can b...

The object tracking by single feature often leads to poor robustness. In this paper, an object tracking algorithm based on multi-features fusion is presented. An adaptive method of choosing object color histogram is presented and the histogram is background weighted in order to get an accurate color model of the object. At meanwhile, then spatiograms feature is applied to obtain spatial layout ...

In this work, we present a novel method for combining predictions of object detection models: weighted boxes fusion. Our algorithm utilizes confidence scores all proposed bounding to constructs the averaged boxes. We tested on several datasets and evaluated it in context Open Images COCO Object Detection tracks, achieving top results these challenges. The source code is publicly available at ht...

Aimed at accurary and real-time object tracking under complex background,an object tracking algorithm based on multi feature fusion is proposed. Feature points tracking is used to reduce the match time and improve the real-time of tracking; To overcome the inaccuracy of a single feature tracking, the object model is presented by the color and texture features. For the traditional "current" stat...

This paper seeks to present a more elegant and general definition of Object Petri Nets than previously. It is more general since it supports transition fusion as well as place fusion. It is more elegant because it captures all the notions of place substitution, transition substitution, place fusion, and transition fusion under the single notion of binding. This is achieved by explicitly support...

In this work, we propose a new integrated framework that addresses the problems of thermal-visible video registration, sensor fusion, and people tracking for far-range videos. The video registration is based on a RANSAC trajectory-totrajectory matching, which estimates an affine transformation matrix that maximizes the overlapping of thermal and visible foreground pixels. Sensor fusion uses the...