## Translation insensitive image similarity in complex wavelet domain (2005)

Venue: | In Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP ’05). IEEE International Conference on |

Citations: | 46 - 5 self |

### BibTeX

@INPROCEEDINGS{Wang05translationinsensitive,

author = {Zhou Wang and Eero P. Simoncelli},

title = {Translation insensitive image similarity in complex wavelet domain},

booktitle = {In Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP ’05). IEEE International Conference on},

year = {2005},

pages = {573--576}

}

### Years of Citing Articles

### OpenURL

### Abstract

We propose a complex wavelet domain image similarity measure, which is simultaneously insensitive to luminance change, contrast change and spatial translation. The key idea is to make use of the fact that these image distortions lead to consistent magnitude and/or phase changes of local wavelet coefficients. Since small scaling and rotation of images can be locally approximated by translation, the proposed measure also shows robustness to spatial scaling and rotation when these geometric distortions are small relative to the size of the wavelet filters. Compared with previous methods, the proposed measure is computationally efficient, and can evaluate the similarity of two images without a precise registration process at the front end. 1.

### Citations

595 | Spatiotemporal Energy Models for the Perception o f M o t i o n . J o u r n a l o f t h e O p t i c a l S o c i e t y o f America, A , v o l . 2 , n o . 2 , p p
- Adelson, Bergan
- 1985
(Show Context)
Citation Context ...; 3) The computation of complex-valued product in visual cortex [10]; 4) The computation of local energy (using sums of squared responses of quadrature-pair filters) by complex cells in visual cortex =-=[11]-=-; and 5) Divisive normalization of filter responses (using summed energy of neighboring filter responses) in both visual and auditory neurons [12]. It is important to realize the limitations of the cu... |

586 | Image quality assessment: from error visibility to structural similarity
- Wang, Bovik, et al.
(Show Context)
Citation Context ...s easy to compute and is mathematically convenient in the context of optimization. However, it has been shown that they perform poorly in image quality assessment and pattern recognition tasks (e.g., =-=[1, 2]-=-). We demonstrate this in Fig. 1, in which Images (b)-(g) have almost the same MSE values with respect to the reference image (a), but exhibit dramatically different perceptual quality as well as reco... |

427 | Shiftable multi-scale transforms
- Simoncelli, Freeman, et al.
- 1992
(Show Context)
Citation Context ... the wavelet filter. 3. TEST To apply the CW-SSIM measure for comparing images, we first decompose the two given images being compared using a complex version [6] of the “steerable pyramid” transform =-=[7]-=- (a type of redundant wavelet transform that avoids aliasing in subbands). We then move a sliding window (of size 7×7) step by step across each wavelet subband. At each step, the CW-SSIM index is calc... |

289 | A parametric texture model based on joint statistics of complex wavelet coefficients
- Portilla, Simoncelli
(Show Context)
Citation Context ...nts does not change the structure of the local image feature. In previous work, similar phase correlation idea had been employed for image alignment [4], feature localization [5], texture description =-=[6]-=- and blur detection [3], but has not been used for image similarity measurement. 2.2. Sensitivity Analysis In all the analysis below, we assume that x corresponds to a reference image and y is an alte... |

177 | Probability distributions of optical flow - Simoncelli, Adelson, et al. - 1991 |

130 | Natural signal statistics and sensory gain control
- Schwartz, Simoncelli
- 2001
(Show Context)
Citation Context ...ature-pair filters) by complex cells in visual cortex [11]; and 5) Divisive normalization of filter responses (using summed energy of neighboring filter responses) in both visual and auditory neurons =-=[12]-=-. It is important to realize the limitations of the current algorithm. First, the CW-SSIM measure does not provide any correspondence information between the pixels of the two images being compared (a... |

126 | Transformation invariance in pattern recognition - tangent distance and tangent propagation
- Simard, Cun, et al.
- 1998
(Show Context)
Citation Context ...s easy to compute and is mathematically convenient in the context of optimization. However, it has been shown that they perform poorly in image quality assessment and pattern recognition tasks (e.g., =-=[1, 2]-=-). We demonstrate this in Fig. 1, in which Images (b)-(g) have almost the same MSE values with respect to the reference image (a), but exhibit dramatically different perceptual quality as well as reco... |

114 |
Feature detection from local energy
- Morrone, Owens
- 1987
(Show Context)
Citation Context ...se shift of all coefficients does not change the structure of the local image feature. In previous work, similar phase correlation idea had been employed for image alignment [4], feature localization =-=[5]-=-, texture description [6] and blur detection [3], but has not been used for image similarity measurement. 2.2. Sensitivity Analysis In all the analysis below, we assume that x corresponds to a referen... |

98 |
Phase-relationships between adjacent simple cells in the visual cortex
- Pollen, Ronner
- 1981
(Show Context)
Citation Context ...he involvement of bandpass visual channels in image pattern recognition tasks [8]; 2) Representation of phase information in primary visual cortex using quadrature pairs of localized bandpass filters =-=[9]-=-; 3) The computation of complex-valued product in visual cortex [10]; 4) The computation of local energy (using sums of squared responses of quadrature-pair filters) by complex cells in visual cortex ... |

49 |
The visual filter mediating letter identification
- Solomon, Pelli
- 1994
(Show Context)
Citation Context ...odels that have been successfully used to account for a variety of biological vision behaviors. These models include: 1) The involvement of bandpass visual channels in image pattern recognition tasks =-=[8]-=-; 2) Representation of phase information in primary visual cortex using quadrature pairs of localized bandpass filters [9]; 3) The computation of complex-valued product in visual cortex [10]; 4) The c... |

22 | Local phase coherence and the perception of blur
- Wang, Simoncelli
- 2004
(Show Context)
Citation Context ... u − p e s jωc(u−p)/s , (1) where s ∈ R + is the scale factor, and p ∈ R is the translation factor. It can be shown that the continuous wavelet transform of a given real signal x(u) can be written as =-=[3]-=-: � ∞ X(s, p) = 1 2π X(ω) −∞ √ s G(s ω − ωc) e jωp dω , (2) where X(ω) and G(ω) are the Fourier transforms of x(u) and g(u), respectively. The discrete wavelet coefficients are sampled versions of the... |

2 |
Stereoscopic depth descrimination in the visual cortex: Neurons ideally suited as disparity detectors
- Ohzawa, DeAngelis, et al.
- 1990
(Show Context)
Citation Context ...tion tasks [8]; 2) Representation of phase information in primary visual cortex using quadrature pairs of localized bandpass filters [9]; 3) The computation of complex-valued product in visual cortex =-=[10]-=-; 4) The computation of local energy (using sums of squared responses of quadrature-pair filters) by complex cells in visual cortex [11]; and 5) Divisive normalization of filter responses (using summe... |

1 |
The phase correlaton image alighment method
- Kuglin, Hines
- 1975
(Show Context)
Citation Context ...ficients. • Consistent phase shift of all coefficients does not change the structure of the local image feature. In previous work, similar phase correlation idea had been employed for image alignment =-=[4]-=-, feature localization [5], texture description [6] and blur detection [3], but has not been used for image similarity measurement. 2.2. Sensitivity Analysis In all the analysis below, we assume that ... |