Moments and Moment Invariants in Pattern Recognition

Flusser, Jan; Zitová, Barbara; Suk, Tomáš

doi:10.1002/9780470684757

Cited by 496 publications

(403 citation statements)

References 12 publications

Supporting

Mentioning

363

Contrasting

Unclassified

Order By: Relevance

“…This is the case of Legendre [1] and Zernike [8]. Another advantage of orthogonal moments is that they commonly have a low computational complexity because, we can evaluate them by using recurrent relations [9]. Hence, if the polynomial basis is orthogonal and satisfies the following condition of orthogonality:…”

Section: Orthogonal Momentsmentioning

confidence: 99%

See 1 more Smart Citation

Experimental Comparison of Orthogonal Moments as Feature Extraction Methods for Character Recognition

Duval

Vega-Pons

Garea

2010

Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

View full text Add to dashboard Cite

Abstract. The selection of a good feature extraction technique is very important in any classification problem. Moments, especially orthogonal moments, seem to be a powerful option in the case of digital image compression, description and recognition. Nowadays, there is a considerable amount of orthogonal moments reported in the literature, each one with some advantages and drawbacks. In this paper, we carry out an experimental comparison of several orthogonal moments for the character recognition problem. Firstly, we compare orthogonal moments with other kinds of feature extraction methods and after that, we compare the different orthogonal moments taking into account different evaluation parameters. Experiments were made by using printed and handwritten digit datasets and the well-known measures: precision, recall and accuracy were used to validate the results. This experimental study corroborates the good performance of orthogonal moments. Besides, more specific results obtained in different kinds of experimentations allow coming to conclusions that could be very useful for the community of image recognition practitioners.

show abstract

Section: Orthogonal Momentsmentioning

confidence: 99%

“…One advantage of this form is that it is invariant with respect to rotation of axes. An example of radial orthogonal functions used for moments construction are Zernike [8], pseudo-Zernike [9] and Generalized pseudo-Zernike [10].…”

Section: Orthogonal Momentsmentioning

confidence: 99%

Experimental Comparison of Orthogonal Moments as Feature Extraction Methods for Character Recognition

Duval

Vega-Pons

Garea

2010

Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

View full text Add to dashboard Cite

show abstract

“…In this respect we have few distinctive groups. The first group of methods belongs to the domain of document processing (Chen et al 2003), the second is logo and trademark recognition in outdoor conditions (Gori et al 2003), such as sport events (Ballan et al 2008), the next one concerns car logo detection and recognition (Dai et al 2009;Psyllos et al 2010;Mao et al 2013), just to name a few.…”

Section: Related Workmentioning

confidence: 99%

“…Then, the dissimilarity measure is computed between sets of line segments rather than sets of points. Gori et al (2003) propose a logo recognition system based on a modified backpropagation neural network. Their edge-backpropagation is a modified learning algorithm derived from the classical backpropagation by addition of a weighting error function in place of the Euclidean norm.…”

Section: Related Workmentioning

confidence: 99%

Hybrid ensemble of classifiers for logo and trademark symbols recognition

Cyganek

2014

Soft Comput

View full text Add to dashboard Cite

The paper presents a hybrid ensemble of diverse classifiers for logo and trademark symbols recognition. The proposed ensemble is composed of four types of different member classifiers. The first one compares color distribution of the logo patterns and is responsible for sifting out images of different color distribution. The second of the classifiers is based on the structural tensor recognition of local phase histograms. A proposed modification in this module consists of tensor computation in the space of the morphological scale-space. Thanks to this, more discriminative histograms describing global shapes are obtained. Next in the chain, is a novel member classifier that joins the Hausdorff distance with the correspondence measure of the log-polar patches computed around the corner points. This sparse classifier allows reliable comparison of even highly deformed patterns. The last member classifier relies on the statistical affine moment invariants which describe global shapes. However, a real advantage is obtained by joining the aforementioned base classifiers into a hybrid ensemble of classifiers, as proposed in this paper. Thanks to this a more accurate response and generalizing properties are obtained at reasonable computational requirements. Experimental results show good recognition accuracy even for the highly deformed logo patterns, as well as fair generalization properties which support human search and logo assessment tasks.

show abstract

“…Invariants of these group actions typically arise to reduce a problem or to decide if two objects, geometric or abstract, are obtained from one another by the action of a group element. [8,9,10,11,13,15,17,39,40,42,43,45,46,52,59] are a few recent references of applications. Both algebraic and differential invariant theories have become in recent years the subject of computational mathematics [13,14,17,40,60].…”

Section: Introductionmentioning

confidence: 99%