Discriminant analysis in correlation similarity measure space

Ma, Yong; Lao, Shihong; Takikawa, Erina; Kawade, Masato

doi:10.1145/1273496.1273569

Cited by 73 publications

(59 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is used to describe similarities between two vectors. This is used in pattern recognition, multivariate statistics ISSN 1948-5433 2016 and data mining (Ma et al, 2007).…”

Section: Correlationmentioning

confidence: 99%

Clustering of OECD Countries Out of Pocket Health Expenditure Time Series Data

Çınaroğlu

2016

RAE

View full text Add to dashboard Cite

Out of pocket health expenditures points out to the payments made by households at the point they receive health services. Frequently these include doctor consultation fees, purchase of medication and hospital bills. In this study hierarchical clustering method was used for classification of 34 countries which are members of OECD (Organization for Economic Cooperation and Development) in terms of out of pocket health expenditures for the years between 1995-2011. Longest common subsequences (LCS), correlation coefficient and Euclidean distance measure was used as a measure of similarity and distance in hierarchical clustering. At the end of the analysis it was found that LCS and Euclidean distance measures were the best for determining clusters. Furthermore, study results led to understand grouping of OECD countries according to health expenditures.

show abstract

“…This is used to describe similarities between two vectors. This is used in pattern recognition, multivariate statistics ISSN 1948-5433 2016 and data mining (Ma et al, 2007).…”

Section: Correlationmentioning

confidence: 99%

Clustering of OECD Countries Out of Pocket Health Expenditure Time Series Data

Çınaroğlu

2016

RAE

View full text Add to dashboard Cite

show abstract

“…It is worth noting that the use of the PCA here is not critical in the sense that any unsupervised subspace learning method such as Kernel PCA, 2DPCA [34], [35], LLE [36], LPP [37] CDA [38], CEA [39], kd-tree, and random projection tree [29] can be used. Denoting F as the feature representation, we measure the similarity score by MCS between two images.…”

Section: Face Verification By Binary-like Representation and Mcsmentioning

confidence: 99%

Face Verification Using the LARK Representation

Seo

Milanfar

2011

IEEE Trans.Inform.Forensic Secur.

140

View full text Add to dashboard Cite

Abstract-We present a novel face representation based on locally adaptive regression kernel (LARK) descriptors [1]. Our LARK descriptor measures a self-similarity based on "signal-induced distance" between a center pixel and surrounding pixels in a local neighborhood. By applying principal component analysis (PCA) and a logistic function to LARK consecutively, we develop a new binary-like face representation which achieves state of the art face verification performance on the challenging benchmark "Labeled Faces in the Wild" (LFW) dataset [2]. In the case where training data are available, we employ one-shot similarity (OSS) [3], [4] based on linear discriminant analysis (LDA) [5]. The proposed approach achieves state of the art performance on both the unsupervised setting and the image restrictive training setting (72.23% and 78.90% verification rates) respectively as a single descriptor representation, with no preprocessing step. As opposed to [4] which combined 30 distances to achieve 85.13%, we achieve comparable performance (85.1%) with only 14 distances while significantly reducing computational complexity.

show abstract

“…Earlier works such as [9,10] have shown that correlation based metrics perform better than the conventional Euclidean and Mahalanobis distances for classification and learning tasks. Motivated by the effectiveness of correlation-based similarity measures, we propose to use cosine similarity for change detection.…”

Section: Technical Detailsmentioning

confidence: 99%

“…To summarize the operation of the overall algorithm, given the reference image and the target image, we first calculate the LSK from both the reference image and the registered target image at all pixel locations. Comparison between LSKs computed from two images is carried out using the cosine similarity measure [9,10]. This step produces a "dissimilarity map" showing the likelihood of dissimilarity between the reference and target images.…”

Section: Introduction and Overviewmentioning

confidence: 99%

A non-parametric approach to automatic change detection in MRI images of the brain

Seo

Milanfar

2009

2009 IEEE International Symposium on Biomedical Imaging: From Nano to Macro

View full text Add to dashboard Cite

We present a novel approach to change detection between two brain MRI scans (reference and target.) The proposed method uses a single modality to find subtle changes; and does not require prior knowledge (learning) of the type of changes to be sought. The method is based on the computation of a local kernel from the reference image, which measures the likeness of a pixel to its surroundings. This kernel is then used as a feature and compared against analogous features from the target image. This comparison is made using cosine similarity. The overall algorithm yields a scalar dissimilarity map (DM), indicating the local statistical likelihood of dissimilarity between the reference and target images. DM values exceeding a threshold then identify meaningful and relevant changes. The proposed method is robust to various challenging conditions including unequal signal strength.

show abstract

Discriminant analysis in correlation similarity measure space

Cited by 73 publications

References 22 publications

Clustering of OECD Countries Out of Pocket Health Expenditure Time Series Data

Clustering of OECD Countries Out of Pocket Health Expenditure Time Series Data

Face Verification Using the LARK Representation

A non-parametric approach to automatic change detection in MRI images of the brain

Contact Info

Product

Resources

About