Number Of Training Images Research Articles

Incremental learning of bidirectional principal components for face recognition

Recently, bidirectional principal component analysis (BDPCA) has been proven to be an efficient tool for pattern recognition and image analysis. Encouraging experimental results have been reported and discussed in the literature. However, BDPCA has to be performed in batch mode, it means that all the training data has to be ready before we calculate the projection matrices. If there are additional samples need to be incorporated into an existing system, it has to be retrained with the whole updated training set. Moreover, the scatter matrices of BDPCA are formulated as the sum of K (samples size) image covariance matrices, this leads to the incremental learning directly on the scatters impossible, thus it presents new challenge for on-line training. In fact, there are two major reasons for building incremental algorithms. The first reason is that in some cases, when the number of training images is very large, the batch algorithm cannot process the entire training set due to large computational or space requirements of the batch approach. The second reason is when the learning algorithm is supposed to operate in a dynamical settings, that all the training data is not given in advance, and new training samples may arrive at any time, and they have to be processed in an on-line manner. Through matricizations of third-order tensor, we successfully transfer the eigenvalue decomposition problem of scatters to the singular value decomposition ( SVD ) of corresponding unfolded matrices, followed by complexity and memory analysis on the novel algorithm. A theoretical clue for selecting suitable dimensionality parameters without losing classification information is also presented in this paper. Experimental results on FERET and CMU PIE (pose, illumination, and expression) databases show that the IBDPCA algorithm gives a close approximation to the BDPCA method, but using less time.

Maximum neighborhood margin criterion in face recognition

Feature extraction is a data analysis technique devoted to removing redundancy and extracting the most discriminative information. In face recognition, feature extractors are normally plagued with small sample size problems, in which the total number of training images is much smaller than the image dimensionality. Recently, an optimized facial feature extractor, maximum marginal criterion (MMC), was proposed. MMC computes an optimized projection by solving the generalized eigenvalue problem in a standard form that is free from inverse matrix operation, and thus it does not suffer from the small sample size problem. However, MMC is essentially a linear projection technique that relies on facial image pixel intensity to compute within- and between-class scatters. The nonlinear nature of faces restricts the discrimination of MMC. Hence, we propose an improved MMC, namely maximum neighborhood margin criterion (MNMC). Unlike MMC, which preserves global geometric structures that do not perfectly describe the underlying face manifold, MNMC seeks a projection that preserves local geometric structures via neighborhood preservation. This objective function leads to the enhancement of classification capability, and this is testified by experimental results. MNMC shows its performance superiority compared to MMC, especially in pose, illumination, and expression (PIE) and face recognition grand challenge (FRGC) databases.

Encoding Probabilistic Brain Atlases Using Bayesian Inference

This paper addresses the problem of creating probabilistic brain atlases from manually labeled training data. Probabilistic atlases are typically constructed by counting the relative frequency of occurrence of labels in corresponding locations across the training images. However, such an "averaging" approach generalizes poorly to unseen cases when the number of training images is limited, and provides no principled way of aligning the training datasets using deformable registration. In this paper, we generalize the generative image model implicitly underlying standard "average" atlases, using mesh-based representations endowed with an explicit deformation model. Bayesian inference is used to infer the optimal model parameters from the training data, leading to a simultaneous group-wise registration and atlas estimation scheme that encompasses standard averaging as a special case. We also use Bayesian inference to compare alternative atlas models in light of the training data, and show how this leads to a data compression problem that is intuitive to interpret and computationally feasible. Using this technique, we automatically determine the optimal amount of spatial blurring, the best deformation field flexibility, and the most compact mesh representation. We demonstrate, using 2-D training datasets, that the resulting models are better at capturing the structure in the training data than conventional probabilistic atlases. We also present experiments of the proposed atlas construction technique in 3-D, and show the resulting atlases' potential in fully-automated, pulse sequence-adaptive segmentation of 36 neuroanatomical structures in brain MRI scans.

Classification of Squamous Cell Carcinoma Based On Color and Textural Features in Microscopic Images of Esophagus Tissues

This paper presents a method for feature extraction using color and texture from microscopic images of esophagus tissues obtained fr om the abnormal regions of human esophagus detected through endoscopy. This method is used for classification of Squamous Cell Carcinoma (SCC) of esophagus, namely, poorly differentiated S CC, moderately differentiated SCC, and well differentiated SCC. Three different color spaces, n amely, HSV, YC bCr, and Lab, are used for color texture analysis to test the classification of SCC of esophagus. The texture features are extracted fr om the luminance channel and the color features are ex tracted from the chrominance channels. The color and textural features are fused to characterize tex ture properties of image. The experimental results show that the classification accuracy of 100% is ob tained using YC bCr color space. Also, the proposed method is robust enough to yield 100% classificatio n rate even with small training/ testing sample in case of poorly differentiated SCC in all the three color spaces. This is a significant result, since t he number of training images is small in most cases an d also the number of testing images of a patient may be small.

Inverse Fisher discriminate criteria for small sample size problem and its application to face recognition

This paper addresses the small sample size problem in linear discriminant analysis, which occurs in face recognition applications. Belhumeur et al. [IEEE Trans. Pattern Anal. Mach. Intell. 19 (7) (1997) 711–720] proposed the FisherFace method. We find out that the FisherFace method might fail since after the PCA transform the corresponding within class covariance matrix can still be singular, this phenomenon is verified with the Yale face database. Hence we propose to use an inverse Fisher criteria. Our method works when the number of training images per class is one. Experiment results suggest that this new approach performs well.

Textural feature based face recognition for single training images

A novel face recognition algorithm using single training face image is proposed. The algorithm is based on textural features extracted using the 2D log Gabor wavelet. These features are encoded into a binary pattern to form a face template which is used for matching. Experimental results show that on the colour FERET database the accuracy of the proposed algorithm is higher than the local feature analysis (LFA) and correlation filter (CF) based face recognition algorithms even when the number of training images is reduced to one. In comparison with recent single training image based face recognition algorithms, the proposed 2D log Gabor wavelet based algorithm shows an improvement of more than 3% in accuracy.

Wavelet packet compression for volume holographic image recognition

An optical image recognition system based on the mechanism of crystal volume holography and wavelet packet compression method is proposed and constructed. Volume holographic associative storage in a photo-refractive crystal has some special properties, which can provide a suitable mechanism to develop an optical correlation system for image recognition. The method of wavelet packet is introduced in this optical system to reduce the number of images stored in the crystal. Through best wavelet packet basis selection, a set of best eigenimages, which will be stored in the crystal as the reference images for recognition, are extracted from a large number of training images. Correlation output between these eigenimages and input ones can be applied for classification and recognition according to the different light intensities of angularly separated beams. Experimental results showed this scheme is practical and can significantly compress the data stored in the crystal while the accurate rate of recognition is still high.

Performance capacity of a simulated annealing bipolar composite filter (SABCF)

Simulated annealing composite filter is a finite element spatial filter. It is trivial to assume that the number of training images to be stored in the SABCF is finite. In this paper, we have investigated the performance capacities for finite element SABCFs. We have shown that performance capacity increases as the number of pixel elements increases. We have also shown that the storage capacity of a SABCF is some what higher if the filter synthesis includes the anti-target training set. In other words, because of the suppression effect on the anti-target set, the SABCF would have a larger capacity as compared with the ones without using the anti-target training set.

Comparison of statistical pattern-recognition algorithms for hybrid processing II Eigenvector-based algorithm

The pattern-recognition algorithms based on eigenvector analysis (group 2) are theoretically and experimentally compared. Group 2 consists of Foley–Sammon (F-S) transform, Hotelling trace criterion (HTC), Fukunaga–Koontz (F-K) transform, linear discriminant function (LDF), and generalized matched filter (GMF) algorithms. It is shown that all eigenvector-based algorithms can be represented in a generalized eigenvector form. However, the calculations of the discriminant vectors are different for different algorithms. Summaries of methods of calculating the discriminant functions for the F-S, HTC, and F-K transforms are provided. Especially for the more practical, underdetermined case, where the number of training images is less than the number of pixels in each image, the calculations usually require the inversion of a large, singular pixel correlation (or covariance) matrix. We suggest solving this problem by finding its pseudoinverse, which requires inverting only the smaller, nonsingular image-correction (or covariance) matrix plus multiplying several nonsingular matrices. We also compare theoretically the classification performance with discriminant functions of the F-S, HTC, and F-K with the LDF and GMF algorithms and the linear-mapping-based algorithms with the eigenvector-based algorithms. Experimentally, we compare the eigenvector-based algorithms, using two sets of image data bases with each image consisting of 64 × 64 pixels.