A supervised pixel-based classifier for identifying the presence of a given set of texture patterns of interest in a complex textured image is described. The proposed technique integrates the outcome of multiple textu...
详细信息
A supervised pixel-based classifier for identifying the presence of a given set of texture patterns of interest in a complex textured image is described. The proposed technique integrates the outcome of multiple texture feature extraction methods belonging to different families. In this way, it yields lower classification rates than previous texture classifiers based on specific families of texture methods. Experimental results with real outdoor images are presented.
This paper proposes a pixel-based texture classifier that integrates multiple texture feature extraction methods in order to identify the regions of an input image that belong to a given set of texture patterns. Exper...
详细信息
This paper proposes a pixel-based texture classifier that integrates multiple texture feature extraction methods in order to identify the regions of an input image that belong to a given set of texture patterns. Experimental results with textured images of outdoor scenes show that the proposed technique yields lower classification errors than widely recognized texture classifiers based on specific families of texture methods.
The shock graph is an emerging shape representation for object recognition, in which a 2-D silhouette is decomposed into a set of qualitative parts, captured in a directed acyclic graph. Although a number of approache...
详细信息
The shockg raph is an emerging shape representation for object recognition, in which a 2-D silhouette is decomposed into a set of qualitative parts, captured in a directed acyclic graph. Although a number of approache...
详细信息
The shock graph is an emerging shape representation for object recognition, in which a 2-D silhouette is decomposed into a set of qualitative parts, captured in a directed acyclic graph. Although a number of approache...
详细信息
The shock graph is an emerging shape representation for object recognition, in which a 2-D silhouette is decomposed into a set of qualitative parts, captured in a directed acyclic graph. Although a number of approaches have been proposed for shock graph matching, these approaches do not address the equally important indexing problem. We extend our previous work in both shock graph matching and hierarchical structure indexing to propose the first unified framework for view-based 3-D object recognition using shock graphs. The heart of the framework is an improved spectral characterization of shock graph structure that not only drives a powerful indexing mechanism (to retrieve similar candidates from a large database), but also drives a matching algorithm that can accommodate noise and occlusion. We describe the components of our system and evaluate its performance using both unoccluded and occluded queries. The large set of recognition trials (over 25,000) from a large database (over 1400 views) represents one of the most ambitious shock graph-based recognition experiments conducted to date.
This work presents an application intended to work as a second reader in Radiology Services for digital mammograms, which makes use of several computervision algorithms. Although the presented prototype basically foc...
详细信息
Presents a method whereby an autonomous mobile robot automatically selects the most informative data from a set of images acquired a priori, using a statistical method termed information sampling. These data could be ...
详细信息
Presents a method whereby an autonomous mobile robot automatically selects the most informative data from a set of images acquired a priori, using a statistical method termed information sampling. These data could be a single pixel or a number scattered throughout an image. This information is then used to build a topological map of the environment. Our sole input data are omnidirectional images obtained from a catadioptric panoramic camera. Experimental results show that by using only the best data the topological position of a robot, visually maneuvering through a simple indoor environment, can easily be determined.
Solid models are the critical data elements in modern computer-aided design (CAD) environments, describing the shape and form of manufactured artifacts. Their growing ubiquity has created new problems in how to effect...
详细信息
ISBN:
(纸本)0769508537
Solid models are the critical data elements in modern computer-aided design (CAD) environments, describing the shape and form of manufactured artifacts. Their growing ubiquity has created new problems in how to effectively manage the many models that are now stored in the digital libraries of large design and manufacturing enterprises. Existing techniques from the engineering literature and from industrial practice, such as group technology, rely on human-supervised encoding and classification, while techniques from the multimedia database and computer graphics/vision communities often ignore the manufacturing attributes most significant in the classification of models. This paper presents our approach to the manufacturing similarly assessment of solid models of mechanical parts based on machining features. Our technical approach is three-fold: (1) perform machining feature extraction to map the solid model to a set of STEP AP 224 machining features; (2) construct a model dependency graph from the set of machining features; and (3) find the nearest neighbors to the query graph using an iterative improvement search across a database of other models. We also present empirical experiments to validate our approach using our testbed, the National Design Repository ( ). The contribution of this research is the first fully automated technique for machining feature-based comparisons of mechanical artifacts. We believe that this work can lead to radical changes in the way in which design data is managed in modern engineering enterprises.
This paper presents an efficient technique for determining a low-cost disassembly sequence suitable to extract a subset of s components from an assembly containing n components, e of which are exterior (e/spl Lt/n). T...
详细信息
This paper presents an efficient technique for determining a low-cost disassembly sequence suitable to extract a subset of s components from an assembly containing n components, e of which are exterior (e/spl Lt/n). The most efficient solution to this so-called geometric selective disassembly problem is the wave propagation algorithm, which is reported to have a computational complexity of O(sn/sup 2/). Instead, the complexity of the proposed algorithm is O(enlogn) when s/spl Lt/n, and O(sn) when s/spl sime/n. Experimental results with synthetic 3D assemblies are presented.
Map building and position estimation are basic tasks in mobile robot navigation with path planning. A method to generate a global map of the vehicle work environment using ultrasonic sensors is developed in this paper...
详细信息
Map building and position estimation are basic tasks in mobile robot navigation with path planning. A method to generate a global map of the vehicle work environment using ultrasonic sensors is developed in this paper. Depending on the physical properties of the walls that form the room where the robot is navigating, sonar sensors show different behaviours. A neural network is utilized to interpret the range readings of ultrasonic sensors in the different environments. A local map composed of squared cells is formed through the neural network that gives the occupancy probabilities for each cell. Finally, a global map is built achieving integration of different views of the environment using Bayes' rule. Results of the method implementation in the construction in a specular environment as well as in rough wall environments are shown in this paper.
暂无评论