Incremental PCA-LDA algorithm

Dagher, Issam

doi:10.1109/cimsa.2010.5611752

Cited by 29 publications

(10 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After extracting the features of hand radiographs, we decompose them into 2-dimensional features by incremental PCA [24] and kernel PCA [25] with different kernel functions. Then, we visualize the 2-dimensional feature distribution.…”

Section: International Journal Of Biomedical Imagingmentioning

confidence: 99%

Fully Automated Bone Age Assessment on Large-Scale Hand X-Ray Dataset

Pan

Zhao

Chen

et al. 2020

International Journal of Biomedical Imaging

View full text Add to dashboard Cite

Bone age assessment (BAA) is an essential topic in the clinical practice of evaluating the biological maturity of children. Because the manual method is time-consuming and prone to observer variability, it is attractive to develop computer-aided and automated methods for BAA. In this paper, we present a fully automatic BAA method. To eliminate noise in a raw X-ray image, we start with using U-Net to precisely segment hand mask image from a raw X-ray image. Even though U-Net can perform the segmentation with high precision, it needs a bigger annotated dataset. To alleviate the annotation burden, we propose to use deep active learning (AL) to select unlabeled data samples with sufficient information intentionally. These samples are given to Oracle for annotation. After that, they are then used for subsequential training. In the beginning, only 300 data are manually annotated and then the improved U-Net within the AL framework can robustly segment all the 12611 images in RSNA dataset. The AL segmentation model achieved a Dice score at 0.95 in the annotated testing set. To optimize the learning process, we employ six off-the-shell deep Convolutional Neural Networks (CNNs) with pretrained weights on ImageNet. We use them to extract features of preprocessed hand images with a transfer learning technique. In the end, a variety of ensemble regression algorithms are applied to perform BAA. Besides, we choose a specific CNN to extract features and explain why we select that CNN. Experimental results show that the proposed approach achieved discrepancy between manual and predicted bone age of about 6.96 and 7.35 months for male and female cohorts, respectively, on the RSNA dataset. These accuracies are comparable to state-of-the-art performance.

show abstract

Section: International Journal Of Biomedical Imagingmentioning

confidence: 99%

Fully Automated Bone Age Assessment on Large-Scale Hand X-Ray Dataset

Pan

Zhao

Chen

et al. 2020

International Journal of Biomedical Imaging

View full text Add to dashboard Cite

show abstract

“…EVD-CCIPCA is an incremental version of popular PCA technique. The main difference between PCA and CCIPCA is that the traditional PCA algorithm computes eigenvectors and eigenvalues for a sample covariance matrix derived from a well known given image data matrix, by solving an eigenvalue system problem [29], whereas EVD based Candid Co-variance Free Incremental PCA methods allow new images and updating the PCA representation each time a new image is introduced [30].…”

Section: Candid Co-variance Free Incremental Pcamentioning

confidence: 99%

Multisensor fusion of visual and thermal images for human face identification using different SVM kernels

Bhowmik

Bhattacharjee

et al. 2012

2012 IEEE Long Island Systems, Applications and Technology Conference (LISAT)

View full text Add to dashboard Cite

In this paper we present a novel method of face identification using different levels of pixel fusion (e.g. ratios for pixel information taken from the visual and thermal images are, 2:3, 1:1, 3:2 and 7:3) and classification of fused images using different kernels of Support Vector Machine (SVM). Visual imagery has been broadly used in face identification systems, but these are very sensitive to illumination changes. This limitation has been overcome by the Infrared (IR) spectrum that provides simpler and more robust solution to boost the identification performance in uncontrolled environments and deliberate attempts to obscure identity. But IR imagery is sensitive to temperature changes in the surrounding environment and variations in the heat patterns of the face and it is opaque to glass. All these facts degrade the face identification efficiency. This drove us to fuse information from both visual and thermal spectra, which have the potential to improve face identification performance as fusion of thermal and visual images provide improved images with more compact information. Once we get fused images those are reduced in dimension using Eigenvalue Decomposition based Candid Co-variance free Incremental Principal Component Analysis (EVD-CCIPCA) and these reduced fused images are classified using the three different kernels of SVM. The three kernels used here are: linear, polynomial and gaussian RBF. SVM is primarily a classifier method that performs classification tasks by constructing hyperplanes in a multidimensional space that separates cases of different class labels. In this paper, we have used multiclass SVM to carry out identification on face images and Quadratic Programming (QP) optimization method to train the SVM. For experiments, IRIS Thermal/Visual Face Database is used. Experimental results show that 97.28% is the highest average success rate achieved on the fused images of 70% visual and 30% thermal images using the linear kernel. However, the highest success rate of 100% is achieved for classes 4 and 10 in several cases.

show abstract

“…x x p N x (17) where ψ is a diagonal covariance matrix whose elements are the corresponding variances of the affine transformation parameters, i.e.,…”

Section: A Dynamic Modelmentioning

confidence: 99%

Robust Visual Tracking via Appearance Modeling and Sparse Representation

Li¹,

Fang-lan²,

Nian³

2014

JCP

View full text Add to dashboard Cite

When appearance variation of object, partial occlusion or illumination change in object images occurs, most existing tracking approaches fail to track the target effectively. To deal with the problem, this paper proposed a robust visual tracking method based on appearance modeling and sparse representation. The proposed method exploits two-dimensional principal component analysis (2DPCA) with sparse representation theory for constructing appearance model. Then tracking is achieved by Bayesian inference framework, in which a particle filter is applied to evaluate the target state sequentially over time. In addition, to make the observation model more robust, the incremental learning algorithm is used to update the template set. Both qualitative and quantitative evaluations on four publicly available benchmark video sequences demonstrate that the proposed visual tracking algorithm performs better than several state-of-the-art algorithms.

show abstract

Incremental PCA-LDA algorithm

Cited by 29 publications

References 27 publications

Fully Automated Bone Age Assessment on Large-Scale Hand X-Ray Dataset

Fully Automated Bone Age Assessment on Large-Scale Hand X-Ray Dataset

Multisensor fusion of visual and thermal images for human face identification using different SVM kernels

Robust Visual Tracking via Appearance Modeling and Sparse Representation

Contact Info

Product

Resources

About