A key element in patternrecognition is the description of shape. The description of shape is facilitated by segmenting the boundary line at so-called critical points - corners (discontinuities in curvature), points o...
详细信息
A key element in patternrecognition is the description of shape. The description of shape is facilitated by segmenting the boundary line at so-called critical points - corners (discontinuities in curvature), points of inflection, and curvature maxima. Additional critical points are intersections and points of tangency. Algorithms are described for extracting such critical points in the presence of noise. An illustration is given showing how the critical points may be used in the development of a shape description system.
Straight-line edge segments are often used as key features in building scene models. This paper presents techniques for extracting edge segments from a digital picture. Results of line extraction on a few outdoor scen...
详细信息
Straight-line edge segments are often used as key features in building scene models. This paper presents techniques for extracting edge segments from a digital picture. Results of line extraction on a few outdoor scenes are also presented.
Sixty-three papers were presented at the conference held at Rensselaer Polytechnic Institute in June, 1977. Areas of coverage were pattern and image analysis, image models, medical applications, speech pattern recogni...
详细信息
Sixty-three papers were presented at the conference held at Rensselaer Polytechnic Institute in June, 1977. Areas of coverage were pattern and image analysis, image models, medical applications, speech patternrecognition, syntactic patternrecognition, industrial applications, shape analysis, scene analysis, multi-imageprocessing, statistical patternrecognition, and character and signature recognition.
To see if nuclear medicine images could be interpreted by computer, lung perfusion images were stored, either on disc, or on magnetic tape as matrices. The matrices were processed, thereafter, by a number of different...
详细信息
To see if nuclear medicine images could be interpreted by computer, lung perfusion images were stored, either on disc, or on magnetic tape as matrices. The matrices were processed, thereafter, by a number of different maneuvers, and an algorithm incorporating the most promising methods was applied to all of the stored images. The images were originally derived from three different scinticameras, with different field sizes, but as long as the orientation of the transmitted imaged was not rotated from what the algorithm was told to expect, the image could be processed. The objective of this project was to write a program to automatically evaluate perfusion lung scans.
A film recording system is described which combines the advantages of the multi-format image systems with the advantages of computer imaging systems. images of improved visual quality are obtained by quantitative cont...
详细信息
A film recording system is described which combines the advantages of the multi-format image systems with the advantages of computer imaging systems. images of improved visual quality are obtained by quantitative control of gray scale and removal of the digital matrix pattern. A method is described for increasing 16 levels of gray to 31 using a double exposure. Formats are defined by the user using a CRT display and interactive software. Formats can consist of curves and text as well as images. A light metering device automatically calibrates the brightness of the display scope.
This paper describes an image dependent two-dimensional nonlinear spatial filter designed to be a resolution preserving textural feature extractor for image data. The textural transform is based on the neighboring gre...
详细信息
This paper describes an image dependent two-dimensional nonlinear spatial filter designed to be a resolution preserving textural feature extractor for image data. The textural transform is based on the neighboring grey tone co-occurrence properties of the image to be transformed. Classification experiments with the textural transform on satellite multispectral scanner imagery over forrested areas show that higher identification accuracy can be achieved when using combined spectral and textural information than when using only spectral information.
The authors consider how to structure the world model so that it will be able to guide the recognitionprocess. For static representation of concepts, visually seen objects, and their relationships, a lattice structur...
详细信息
The authors consider how to structure the world model so that it will be able to guide the recognitionprocess. For static representation of concepts, visually seen objects, and their relationships, a lattice structure has been used. The lattice structure is felt to embody certain natural orderings among concepts, such as the generalization and complexity aspects. In addition to the static representation, there has been introduced a dynamic mechanism called conceptual focusing and defocusing. This process enables one to organize the search for certain objects in more efficient and meaningful way.
The system consists of a digital image scanner used for locating the lung fields and for texture measurements as well as an automated film transport and Fraunhofer diffraction pattern sampling unit. A minicomputer is ...
详细信息
The system consists of a digital image scanner used for locating the lung fields and for texture measurements as well as an automated film transport and Fraunhofer diffraction pattern sampling unit. A minicomputer is used to control the image scanning, position the film in the optical system and compute discriminate functions and probability estimates for the classification decisions. Results for a normal-abnormal classification are presented using both digital and optical texture measurements. These results correlate well with previous feasibility studies and with physician readings anindicate that an alternative to visual classification is possible through the use of automated measurements.
Feasibility of identification of three-dimensional objects from their silhouettes by a sequential matching algorithm is studied. A known object is sequentially moved in the parameter space, and synthetically projected...
详细信息
Feasibility of identification of three-dimensional objects from their silhouettes by a sequential matching algorithm is studied. A known object is sequentially moved in the parameter space, and synthetically projected until its silhouette matches that of the unknown silhouette. The method appears to work well but requires substantial numerical computation, since many local minima exist. It may be more suited for tracking trajectories of moving objects where local minima would provide estimation of current parameters of the object, and where excessive processing may not be needed.
Efficient data structures for picture processing are developed on the basis of graph theoretic concepts. Picture segmentation is achieved by a directed split-and-merge algorithm and subsequent analysis is performed on...
详细信息
Efficient data structures for picture processing are developed on the basis of graph theoretic concepts. Picture segmentation is achieved by a directed split-and-merge algorithm and subsequent analysis is performed on the basis of the region adjacency graph (RAG). It is shown that multiply connected regions (i. e. having holes) correspond to cutnodes of this graph. The latter can be identified easily by using the Hopcroft-Tarjan algorithm.
暂无评论