The neocognitron is a neural model which can recognize a pattern even when it is shifted in position, changed in size, or distorted. However, the complexity of its structure and its algorithm makes it difficult to tun...
详细信息
The neocognitron is a neural model which can recognize a pattern even when it is shifted in position, changed in size, or distorted. However, the complexity of its structure and its algorithm makes it difficult to tune its parameters and to specify the detail of its structure. The characteristics of this model are measured as the relationship between the degree of distortion of the input and the tolerance of the model. Computational experiments show that the pyramid has a suitable structure for implementing a neocognitron, provided that the input has a relatively low spatial frequency.< >
Many objects that appears in digital images are bounded by straight lines. Curves are often detected as series of edges. The detection of the straightness of those edge sequences is a major problem in image interpreta...
详细信息
Many objects that appears in digital images are bounded by straight lines. Curves are often detected as series of edges. The detection of the straightness of those edge sequences is a major problem in image interpretation. In this paper the RULI chain code is used for encoding geometric configurations. Rosenfeld ( IEEE Trans. Comput. C-23 , 1974, 1264–1269) and Ronse ( Pattern Recognit. Lett. 3 , 1985, 323–326) used the “chord property” to identify straight lines. While this test takes O(n ∗ ∗ 2) steps, the algorithm discussed in this paper only takes O ( n ) steps, where n is the number of code elements. Furthermore, the algorithm calculates the interval that delimits the slope of the straight line.
Three-dimensional restitution of images with unconventional imaging geometry can be performed in an operational mode using existing commercial analytical plotters without any extra investments in hardware. The paper p...
Three-dimensional restitution of images with unconventional imaging geometry can be performed in an operational mode using existing commercial analytical plotters without any extra investments in hardware. The paper presents two programme systems developed for the Kern DSR analytical stereo-plotter for mapping with spaceborne stereo-imagery, one for images acquired by the SPOT HRV sensor, the other one for stereo-mensuration of SAR images. Due to the movement of the satellite platforms the software requires appropriate algorithms for the consideration of time-varying orientation parameters like sensor position, sensor attitude or others. The performance of the programme systems is demonstrated with practical results derived from Space Shuttle SIR-B imagery as well as from SPOT images showing realistically achievable accuracies of a few pixels in planimetry.
Spaceborne stereo-imagery can serve two major purposes: relief mapping and extraction of thematic information. Not only camera-acquired photogrammetric data of conventional and super-large format, but also hand-held p...
暂无评论