Class-specific material categorisation

Caputo, Barbara; Hayman, Eric; Mallikarjuna, P.

doi:10.1109/iccv.2005.54

Cited by 251 publications

(233 citation statements)

References 15 publications

Supporting

Mentioning

231

Contrasting

Order By: Relevance

“…Caputo et al [14] report 71.0% for their 3-scale LBP and non-linear chi-squared RBF kernel based SVM classifier. In comparison we use linear classifiers which are not only fast to train but also need only a vector dot product at test time (c.f .…”

Section: Comparison With Existing Methodsmentioning

confidence: 99%

“…All these variations make it an extremely challenging dataset. We use the standard protocol [11,14] and report the average performance over the 4 runs, where every time all images of one sample are taken for test while the images of the remaining 3 samples are used for training.…”

Section: Kth Tips 2a Datasetmentioning

confidence: 99%

“…Hence, the coding of vectors is intrinsically adapted to any classification task and the computations involved remain very simple despite the strengths. We validate our approach by extensive experiments on four challenging datasets: (i) Brodatz 32 texture dataset [12,13], (ii) KTH TIPS 2a materials dataset [14], (iii) Japanese female facial expressions dataset [15], and (iv) Labeled Faces in the Wild (LFW) dataset [16], and show that using higher-order statistics gives a more expressive description and leads to state-of-the-art performance.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Local Higher-Order Statistics (LHS) for Texture Categorization and Facial Analysis

Sharma

Hussain

Jurie

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper proposes a new image representation for texture categorization and facial analysis, relying on the use of higher-order local differential statistics as features. In contrast with models based on the global structure of textures and faces, it has been shown recently that small local pixel pattern distributions can be highly discriminative. Motivated by such works, the proposed model employs higher-order statistics of local non-binarized pixel patterns for the image description. Hence, in addition to being remarkably simple, it requires neither any user specified quantization of the space (of pixel patterns) nor any heuristics for discarding low occupancy volumes of the space. This leads to a more expressive representation which, when combined with discriminative SVM classifier, consistently achieves state-of-the-art performance on challenging texture and facial analysis datasets outperforming contemporary methods (with similar powerful classifiers).

show abstract

Section: Comparison With Existing Methodsmentioning

confidence: 99%

Section: Kth Tips 2a Datasetmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Local Higher-Order Statistics (LHS) for Texture Categorization and Facial Analysis

Sharma

Hussain

Jurie

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…We evaluate the performance of the proposed method on three comprehensive databases by classifying textures in different scenarios: Outex database [17] for rotation invariant texture classification, KTH-TIPS2 database [3] for material categorization and Columbia-Utrecht (CUReT) database [6] for classification under different views and illuminations, as seen in Figure 3. LCP is also compared with several state-of-the-art texture classification approaches on all these databases.…”

Section: Methodsmentioning

confidence: 99%

“…The proposed method is also assessed on the KTH-TIPS2 database [3]. This database contains 11 different classes of texture materials, each class has 4 different samples, and each sample was imaged at 9 different scales and 12 different illumination and pose conditions.…”

Section: Materials Categorizationmentioning

confidence: 99%

Texture Classification using a Linear Configuration Model based Descriptor

Guo¹,

Zhao²,

Pietikäinen³

2011

Procedings of the British Machine Vision Conference 2011

View full text Add to dashboard Cite

We investigate rotation invariant image description and develop a linear model based descriptor namely MiC, which is suited to modeling microscopic configuration of images. To explore multi-channel discriminative information of both the microscopic configuration and local structures, the feature extraction process is formulated as an unsupervised framework that consists of: 1) the configuration model to encode image microscopic configuration; and 2) local patterns to describe local structural information. In this way, images are represented by a novel feature: local configuration pattern (LCP). We evaluate the performance of the proposed method by classifying textures present in three challenging texture databases: Outex_TC_00012, KTH-TIPS2 and Columbia-Utrecht (CUReT). The encouraging results show that LCPs is highly discriminative.

show abstract