2D Shape Recognition Using Information Theoretic Kernels

Bicego, Manuele; Martins, André F. T.; Murino, Vittorio; Aguiar, Pedro M. Q.; Figueiredo, Mário A. T.

doi:10.1109/icpr.2010.15

Cited by 6 publications

(7 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…cyclic dist [21] 0.78 K-NN + cyclic string edit dist [22] 0.74 1-NN + mBm-based features [23] 0.77 1-NN + HMM-based distance [23] 0.74 1-NN + IT kernels on n-grams [24] 0.81 1-NN + BLOSUM (local alignment) [8] 0.83…”

Section: B Resultsmentioning

confidence: 99%

“…One shape per class is shown. (6,2), which is the lowest pair allowed by many tools publicly available 2 .…”

Section: A Parameters Setting and Experimental Detailsmentioning

confidence: 99%

“…Several approaches have been proposed in the past (see for example the reviews [1], [2]), and many of them are based on the analysis of the boundary: very often, the 2D shape is encoded by its contour, which proved to be an effective and perceptually reasonable choice in many applications. Different techniques exhibit different characteristics: robustness to noise and occlusions, invariance to translation, rotation, and scale, computational requirements, and accuracy (see [3], [4], [5], [6] and references therein).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

S-BLOSUM: Classification of 2D Shapes with Biological Sequence Alignment

Lovato

Milanese

Centomo

et al. 2014

2014 22nd International Conference on Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Recent works investigated the possibility to design solutions for pattern recognition problems by exploiting the huge amount of work done in bioinformatics. If the pattern recognition problem is cast in biological terms, then a huge range of algorithms, exploitable for classification, detection, visualization, etc. can be effectively borrowed. In this paper, we exploit biological sequence alignment tools to classify 2D shapes, tailoring the biological parameters of these tools to account for the different semantic of the 2D shape scenario. In particular, we propose a novel substitution matrix, which is the crucial parameter determining the sequence alignment solution. The new matrix, called S-BLOSUM, learns the rates of matches/mismatches in conserved portions of shapes belonging to the same category, and incorporates prior knowledge on the chosen representation for the 2D shape. On one hand, the experimental evaluation showed that the S-BLOSUM provides a significant improvement over the biological counterpart (BLOSUM); on the other hand, classification results prove that our approach is competitive with respect to the state of the art.

show abstract

Section: B Resultsmentioning

confidence: 99%

“…One shape per class is shown. (6,2), which is the lowest pair allowed by many tools publicly available 2 .…”

Section: A Parameters Setting and Experimental Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

S-BLOSUM: Classification of 2D Shapes with Biological Sequence Alignment

Lovato

Milanese

Centomo

et al. 2014

2014 22nd International Conference on Pattern Recognition

Self Cite

View full text Add to dashboard Cite

show abstract

“…As a final comment, in Table 2 we reported some other recent results from the state of the art on the same datasets. Many different approaches have been tested on the Chicken dataset, using simple as well complicated classifiers (see for example comparisons reported in [20,21]): in Table 2(a) we reported only those based on nearest neighbour rules -taken from [20]. Even if in some cases different experimental protocols have been employed, it seems evident that the proposed approach represents a promising alternative to classic as well as to advanced schemes.…”

Section: Resultsmentioning

confidence: 99%

2D Shapes Classification Using BLAST

Lovato

Bicego

2012

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. This paper presents a novel 2D shape classification approach, which exploits in this context the huge amount of work carried out by bioinformaticians in the biological sequence analysis research field. In particular, in the approach presented here, we propose to encode shapes as biological sequences, employing the widely known sequence alignment tool called BLAST (Basic Local Alignment Search Tool) to devise a similarity score, used in a nearest neighbour scenario. Obtained results on standard datasets show the feasibility of the proposed approach.

show abstract

“…Chen, R. Feris, and M. Turk [2] proposed an efficient technique to find only the similar portions of two shapes and calculate the likeness of the shapes using only similar portions. Here the main contribution is to suggest a local shape identifying technique based on the Smith-Waterman algorithm [14] to efficiently find common portions of two shapes without needing exhaustive search for all likely parts.…”

Section: Literature Surveymentioning

confidence: 99%

Adopting Biological Sequence Alignment Tools in 2D Shape Recognition

VP¹

2015

IJIRSET

View full text Add to dashboard Cite

ABSTRACT:Recognition of two dimensional shapes is without doubts an significant and still open research part in computer vision and pattern recognition .Adopting biological alignment tools into this problem, sequence alignment in biological field is already investigated by different techniques. Human vision reveals that the shape feature is a significant perception to differentiate different objects. This stimulates researchers to find consistent and efficient shape features for discerning objects by machine vision. Many methods have been suggested in the past , many of them based one on features extracted from the boundary: actually, shape contours have shown to be very sensitive in many perspectives. In the proposed method, we encrypt shapes as biological sequences, using standard and well recognized sequence alignment tools to invent a similarity score, finally used in a nearest neighbour situation.

show abstract

2D Shape Recognition Using Information Theoretic Kernels

Abstract: In this paper, a novel approach for contour-based 2D shape recognition is proposed, using a recently introduced class of information theoretic kernels.

Cited by 6 publications

References 13 publications

S-BLOSUM: Classification of 2D Shapes with Biological Sequence Alignment

S-BLOSUM: Classification of 2D Shapes with Biological Sequence Alignment

2D Shapes Classification Using BLAST

Adopting Biological Sequence Alignment Tools in 2D Shape Recognition

Contact Info

Product

Resources

About