This study emphasises the necessity of a global comprehension of a map to afford the exhaustive extraction of the different themes. Spatial informations in maps are encoded by structures and arrangements of graphics p...
详细信息
This study emphasises the necessity of a global comprehension of a map to afford the exhaustive extraction of the different themes. Spatial informations in maps are encoded by structures and arrangements of graphics primitives. A complete method as well as a preliminary study on the map to interpret are described. A scenario of actions leading to the extraction of themes is then created. These actions are divided into three levels according to their abstraction level in the process;low-level extractions, intra-layer handling and inter-layer operators.
An approach to determining the textural attributes of an image on a pixel-by-pixel basis is presented. The technique is based on modelling the spatial interactions within the image using a Gaussian Markov Random Field...
详细信息
An approach to determining the textural attributes of an image on a pixel-by-pixel basis is presented. The technique is based on modelling the spatial interactions within the image using a Gaussian Markov Random Field. The method is applied to the classification of homogeneous texture images, artificial texture collages and a composite-texture SAR terrain image. The capabilities and limitations of the technique are considered in the light of these classification experiments.
This paper presents an adaptive compression scheme for still images. According to the spatial activity of each block of the image, a different number of DCT coefficients are calculated. A lattice vector quantizer, whe...
详细信息
This paper presents an adaptive compression scheme for still images. According to the spatial activity of each block of the image, a different number of DCT coefficients are calculated. A lattice vector quantizer, where pyramids with different radius and dimensionality are used for the vector quantization of the coefficients of each block. The proposed algorithm has low computational complexity and results in high quality image reconstruction at low bit-rates per pixel.
This work investigates an approach to separate time-variant and time-invariant regions in sector-scan sonar image sequence in both time domain and frequency domain. Combining time domain and frequency domain methods, ...
详细信息
This work investigates an approach to separate time-variant and time-invariant regions in sector-scan sonar image sequence in both time domain and frequency domain. Combining time domain and frequency domain methods, the merging region between moving and static objects in a sequence of sonar images can be segmented. This method is useful not only for image segmentation but also for object tracking.
BORG is a knowledge-based system that automates the execution of image segmentation processes. Given an image and a set of segmentation tasks, BORG generates, performs, and optimises the processing.of classical segmen...
详细信息
BORG is a knowledge-based system that automates the execution of image segmentation processes. Given an image and a set of segmentation tasks, BORG generates, performs, and optimises the processing.of classical segmentation operators thereby producing images that satisfy these tasks. The design principles for building image segmentation applications based on the generation of a graph of operators capable of dealing with the whole set of images belonging to some given class are described. A model for building this graph as well as the system architecture are presented. Finally, an application dealing with biomedical images is discussed.
In order to develop a shape and decorative pattern database of old pottery and ceramics objects,it is neccessary to obtain the surface pattern on the 3- dimensional shape of the object. This paper describes the recove...
详细信息
Solving a symbolic description of aspects or contents of the image involves labeling a set of objects such that specific constraints are satisfied. This paper focuses on knowledge representation based on semantic grap...
详细信息
Solving a symbolic description of aspects or contents of the image involves labeling a set of objects such that specific constraints are satisfied. This paper focuses on knowledge representation based on semantic graphs. In this context, control algorithms based on arc consistency are considered, for example, 'true three dimensional images. This algorithm has been tested on twenty N.M.R. cerebral images. The test show the reliability of this method to identify the three-dimensional objects which are inside a set of image.
This paper describes a prototype module, which is an image acquisition and processing.pipeline implemented in a micro-package. The module converts incident light directly to processed digital video data. The system is...
详细信息
This paper describes a prototype module, which is an image acquisition and processing.pipeline implemented in a micro-package. The module converts incident light directly to processed digital video data. The system is packaged using a novel three-dimensional technology called MCM-V (Multichip Module-Vertical), Val (1). This technology stacks integrated circuits in the z-axis to produce dense systems. The module is fully programmable and can be used in a wide range of imageprocessing.tasks. The results presented here form the first stage in the development of high performance ultra-miniature camera and imageprocessing.systems.
The real-time measurement of traffic queue parameters are required in many traffic situations such as accident and congestion monitoring and adjusting the timings of the traffic lights. So far the reported image proce...
详细信息
The real-time measurement of traffic queue parameters are required in many traffic situations such as accident and congestion monitoring and adjusting the timings of the traffic lights. So far the reported imageprocessing.methods have been targeted for measuring simple traffic parameters. In this paper we describe imageprocessing.techniques together with the results to measure the queue traffic parameters in real-time. The proposed queue detection algorithm consists of a motion detection and vehicle detection operation, both based on extracting edges of the scene. The results show that the proposed algorithms are able to measure various queue parameters such as queue detection, length of the queue, period of the occurrence of the queue, slope of the queue etc.
Three theories have been developed concerning the convergence of light intensity towards an object edge and a new technique, unique to the single edge characteristic of the above theories, is established to enhance lo...
详细信息
Three theories have been developed concerning the convergence of light intensity towards an object edge and a new technique, unique to the single edge characteristic of the above theories, is established to enhance local details in video conference pictures for efficient coding. This convergence method or mode of 'sliding' along the edge is mathematically modelled for contour detection. It has the merits of a single edge characteristic, the depiction of key edge points, the constant velocity criterion of the model beneficial to local movement, the correspondence of the theory with the predicted edge points and the potential to develop higher-order models for further investigation.
暂无评论