著者

清水 創太 Burdick Joel W.

出版者

一般社団法人日本機械学会

雑誌

日本機械学會論文集. C編 (ISSN:03875024)

巻号頁・発行日

vol.73, no.733, pp.2591-2596, 2007-09-25

被引用文献数

1

---

This paper aims at acquiring robust feature for rotation-, scale-, and translation-invariant image matching from a space-variant image by a fovea sensor. A proposed model of eccentric compensator corrects deformation in a log-polar image when the fovea sensor is not centered at a target image, that is, eccentricity exists. An image simulator in discrete space implements this model by its geometrical formulation. This paper also proposes Unreliable Feature Omission (UFO) using Discrete Wavelet Transform. UFO reduces local high frequency noise appeared in the space-variant image when the eccentricity changes. It discards coefficients when they are regarded as unreliable, based on digitized errors of the input image by the fovea sensor. The first simulation estimates the compensator by comparing with other polar images. This result shows the compensator works well and its root mean square error (RMSE) changes only by up to 2.54 [%], in condition of the eccentricity within 34.08 [°]. The second simulation shows UFO works well for the log-polar image remapped by the eccentricity compensator, when white Gaussian noise (WGN) is added. The result by Daubechies (7, 9) biorthogonal wavelet shows UFO reduces the RMSE by up to 0.40 [%] even if the WGN is not added, when the eccentricity is within 34.08 [°].

著者

清水 創太

出版者

一般社団法人日本機械学会

雑誌

日本機械学會論文集. C編 (ISSN:03875024)

巻号頁・発行日

vol.73, no.731, pp.2095-2100, 2007-07-25

被引用文献数

4

---

This paper proposes a model of wide-angle spase-variant image that provides a guide for designing fovea sensor. First, Advanced Wide-Angle Foveated(AdWAF) model is formulized taking all-purpose use into account. This proposed model uses both linear coordinates and logarithmic coordinates in both planar projection and spherical projection. Thus, this model divides its 120-degree wide-angle field of view into 4 areas, such that it can represent images by various types of lens, flexibly. The first simulation compares with other lens models, in terms of image height and resolution. The result shows the AdWAF model can reduce image data by 13.5 percent, comparing with a log-polar lens model, when both have the same resolution in the fovea area. The second simulation compares with other foveation models used for the existing wide-angle foveated lens, logpolar chip and vision system. The third simulation estimates scale-invariant property by comparing with the Kuniyoshi's lens and the log-polar lens. The AdWAF model gives its planar logarithmic part complete scale-invariant property, while the Kuniyoshi's lens has about 7.6-percent error at most in its spherical logarithmic part. Moreover, the AdWAF image is extracted from an input image by wide-angle foveated(WAF) lens, based on the proposed model.