Segmentation of magnetic resonance images using fuzzy algorithms for learning vector quantization

This paper evaluates a segmentation technique for magnetic resonance (MR) images of the brain based on fuzzy algorithms for learning vector quantization (FALVQ). These algorithms perform vector quantization by updating all prototypes of a competitive network through an unsupervised learning process. Segmentation of MR images is formulated as an unsupervised vector quantization process, where the local values of different relaxation parameters form the feature vectors which are represented by a relatively small set of prototypes. The experiments evaluate a variety of FALVQ algorithms in terms of their ability to identify different tissues and discriminate between normal tissues and abnormalities.     

Image compression using wavelet transform and multiresolutiondecomposition

Schemes for image compression of black-and-white images based on the wavelet transform are presented. The multiresolution nature of the discrete wavelet transform is proven as a powerful tool to represent images decomposed along the vertical and horizontal directions using the pyramidal multiresolution scheme. The wavelet transform decomposes the image into a set of subimages called shapes with different resolutions corresponding to different frequency bands. Hence, different allocations are tested, assuming that details at high resolution and diagonal directions are less visible to the human eye. The resultant coefficients are vector quantized (VQ) using the LGB algorithm. By using an error correction method that approximates the reconstructed coefficients quantization error, we minimize distortion for a given compression rate at low computational cost. Several compression techniques are tested. In the first experiment, several 512x512 images are trained together and common table codes created. Using these tables, the training sequence black-and-white images achieve a compression ratio of 60-65 and a PSNR of 30-33. To investigate the compression on images not part of the training set, many 480x480 images of uncalibrated faces are trained together and yield global tables code. Images of faces outside the training set are compressed and reconstructed using the resulting tables. The compression ratio is 40; PSNRs are 30-36. Images from the training set have similar compression values and quality. Finally, another compression method based on the end vector bit allocation is examined.     

Multiresolution Vector Quantization for Video Coding

In this work, we propose a coding technique that is based on the generalized block prediction of the multiresolution subband decomposition of motion compensated difference image frames. A segmentation mask is used to distinguish between the regions where motion compensation was effective and those regions where the motion model did not succeed. The difference image is decomposed into a multiresolution pyramid of subbands where the highest resolution subbands are divided into two regions, based on the information given by the segmentation mask. Only the coefficients of the regions corresponding to the motion model failure are considered in the highest resolution subbands. The remaining coefficients are coded using a multiresolution vector quantization scheme that exploits inter-band non-linear redundancy. In particular, blocks in one subimage are predicted from blocks of the adjacent lower resolution subimage with the same orientation. This set of blocks plays the role of a codebook built from coefficients inside the subband decomposition itself. Whenever the inter-band prediction does not give satisfactory results with respect to a target quality, the block coefficients are quantized using a lattice vector quantizer for a Laplacian source.     

Use of nonlinear principal component analysis and vectorquantization for image coding

The nonlinear principal component analysis (NLPCA) method is combined with vector quantization for the coding of images. The NLPCA is realized using the backpropagation neural network (NN), while vector quantization is performed using the learning vector quantizer (LVQ) NN. The effects of quantization in the quality of the reconstructed images are then compensated by using a novel codebook vector optimization procedure.     

Image coding by block prediction of multiresolution subimages

The redundancy of the multiresolution representation has been clearly demonstrated in the case of fractal images, but it has not been fully recognized and exploited for general images. Fractal block coders have exploited the self-similarity among blocks in images. We devise an image coder in which the causal similarity among blocks of different subbands in a multiresolution decomposition of the image is exploited. In a pyramid subband decomposition, the image is decomposed into a set of subbands that are localized in scale, orientation, and space. The proposed coding scheme consists of predicting blocks in one subimage from blocks in lower resolution subbands with the same orientation. Although our prediction maps are of the same kind of those used in fractal block coders, which are based on an iterative mapping scheme, our coding technique does not impose any contractivity constraint on the block maps. This makes the decoding procedure very simple and allows a direct evaluation of the mean squared error (MSE) between the original and the reconstructed image at coding time. More importantly, we show that the subband pyramid acts as an automatic block classifier, thus making the block search simpler and the block matching more effective. These advantages are confirmed by the experimental results, which show that the performance of our scheme is superior for both visual quality and MSE to that obtainable with standard fractal block coders and also to that of other popular image coders such as JPEG.     

Image segmentation using hidden Markov Gauss mixture models.

Image segmentation is an important tool in image processing and can serve as an efficient front end to sophisticated algorithms and thereby simplify subsequent processing. We develop a multiclass image segmentation method using hidden Markov Gauss mixture models (HMGMMs) and provide examples of segmentation of aerial images and textures. HMGMMs incorporate supervised learning, fitting the observation probability distribution given each class by a Gauss mixture estimated using vector quantization with a minimum discrimination information (MDI) distortion. We formulate the image segmentation problem using a maximum a posteriori criteria and find the hidden states that maximize the posterior density given the observation. We estimate both the hidden Markov parameter and hidden states using a stochastic expectation-maximization algorithm. Our results demonstrate that HMGMM provides better classification in terms of Bayes risk and spatial homogeneity of the classified objects than do several popular methods, including classification and regression trees, learning vector quantization, causal hidden Markov models (HMMs), and multiresolution HMMs. The computational load of HMGMM is similar to that of the causal HMM.     

Predictive subband image coding with wavelet transform

Wavelet transform can decompose images into various multiresolution subbands. In these subbands the correlation exists. A novel technique for image coding by taking advantage of the correlation is addressed. It is based on predictive edge detection from the LL band of the lowest resolution level to predict the edge in the LH, HL and HH bands in the higher resolution level. If the coefficient is predicted as an edge it is preserved; otherwise, it is discarded. In the decoder, the location of the preserved coefficients can also be found as in the encoder. Therefore, no overhead is needed. Instead of complex vector quantization, which is commonly used in subband image coding for high compression ratio, simple scalar quantization is used to code the remaining coefficients and achieves very good results.     

Subband Finite-State Vector Quantization for Still Image Coding

Subband coding (SBC) with vector quantization (VQ) has been shown to be an effective method for coding images at low bit rates. The basic idea of subband coding is to split up the frequency band of the signal and then to encode the subbands. Reconstruction is performed by decoding and merging the interpolated subband images. In VQ, the image to be encoded is first processed to yield a set of vectors. The input vectors are individually quantized to the closest codewords in the codebook. In this paper, we propose a new subband finite-state vector quantization (SBC-FSVQ) scheme that combines the SBC and the FSVQ. The frequency band decomposition of an image is carried out by means of 2D separable quadrature mirror filters (QMFs). In our coding scheme, we split the image spectrum into sixteen equally sized subbands. The FSVQ is used to improve the performance by using the correlations of the neighboring samples in the same subband. Thus, our SBC-FSVQ scheme not only has the advantages of the SBC-VQ scheme but also reduces the bit rate and improves the image quality. Experimental results are given and comparisons are made using our new schemes and some other coding techniques. Our technique yields good PSNR performance, for images both inside and outside a training set of five 512 × 512 images. In the experiments, it is found that our SBC-FSVQ scheme achieves the best PSNR performance at nearly the same bit rate.     

Design of vector quantizer for image compression using self-organizing feature map and surface fitting

We propose a new scheme of designing a vector quantizer for image compression. First, a set of codevectors is generated using the self-organizing feature map algorithm. Then, the set of blocks associated with each code vector is modeled by a cubic surface for better perceptual fidelity of the reconstructed images. Mean-removed vectors from a set of training images is used for the construction of a generic codebook. Further, Huffman coding of the indices generated by the encoder and the difference-coded mean values of the blocks are used to achieve better compression ratio. We proposed two indices for quantitative assessment of the psychovisual quality (blocking effect) of the reconstructed image. Our experiments on several training and test images demonstrate that the proposed scheme can produce reconstructed images of good quality while achieving compression at low bit rates. Index Terms-Cubic surface fitting, generic codebook, image compression, self-organizing feature map, vector quantization.     

Subband coding of images using vector quantization

A novel two-dimensional subband coding technique is presented that can be applied to images as well as speech. A frequency-band decomposition of the image is carried out by means of 2D separable quadrature mirror filters, which split the image spectrum into 16 equal-rate subbands. These 16 parallel subband signals are regarded as a 16-dimensional vector source and coded as such using vector quantization. In the asymptotic case of high bit rates, a theoretical analysis yields that a lower bound to the gain is attainable by choosing this approach over scalar quantization of each subband with an optimal bit allocation. It is shown that vector quantization in this scheme has several advantages over coding the subbands separately. Experimental results are given, and it is shown the scheme has a performance that is comparable to that of more complex coding techniques     

Finite-state vector quantization by exploiting interband andintraband correlations for subband image coding

Subband coding (SBC) with vector quantization (VQ) has been shown to be an effective method for coding images at low bit rates. We split the image spectrum into seven nonuniform subbands. Threshold vector quantization (TVQ) and finite state vector quantization (FSVQ) methods are employed in coding the subband images by exploiting interband and intraband correlations. Our new SBC-FSVQ schemes have the advantages of the subband-VQ scheme while reducing the bit rate and improving the image quality. Experimental results are given and comparisons are made using our new scheme and some other coding techniques. In the experiments, it is found that SBC-FSVQ schemes achieve the best peak signal-to-noise ratio (PSNR) performance when compared to other methods at the same bit rate.     

A novel fingerprint image compression technique using wavelets packets and pyramid lattice vector quantization

A novel compression algorithm for fingerprint images is introduced. Using wavelet packets and lattice vector quantization , a new vector quantization scheme based on an accurate model for the distribution of the wavelet coefficients is presented. The model is based on the generalized Gaussian distribution. We also discuss a new method for determining the largest radius of the lattice used and its scaling factor , for both uniform and piecewise-uniform pyramidal lattices. The proposed algorithms aim at achieving the best rate-distortion function by adapting to the characteristics of the subimages. In the proposed optimization algorithm, no assumptions about the lattice parameters are made, and no training and multi-quantizing are required. We also show that the wedge region problem encountered with sharply distributed random sources is resolved in the proposed algorithm. The proposed algorithms adapt to variability in input images and to specified bit rates. Compared to other available image compression algorithms, the proposed algorithms result in higher quality reconstructed images for identical bit rates.     

A multiresolution approach for rotation invariant texture image retrieval with orthogonal polynomials model

In this paper, a simple and an efficient Content Based Image Retrieval which is based on orthogonal polynomials model is presented. This model is built with a set of carefully chosen orthogonal polynomials and is used to extract the low level texture features present in the image under analysis. The orthogonal polynomials model coefficients are reordered into multiresolution subband like structure. Simple statistical and perceptual properties are derived from the subband coefficients to represent the texture features and these features form a feature vector. The efficiency of the proposed feature vector extraction for texture image retrieval is experimented on the standard Brodatz and MIT’s VisTex texture database images with the Canberra distance measure. The proposed method is compared with other existing retrieval schemes such as Discrete Cosine Transformation (DCT) based multiresolution subbands, Gabor wavelet and Contourlet Transform based retrieval schemes and is found to outperform the existing schemes with less computational cost.     

采用方向树结构矢量组合的小波图像分类矢量量化

该文提出了采用方向树结构矢量组合对小波图像进行分类矢量量化的新方法。该方法的矢量构成结合了子带系数的方向性,充分利用了子带系数的带间和带内的相关性,按能量和活跃度进行两级分类,降低了类中矢量的内部离散度,分类信息占用比特数少,并采用了基于人眼视觉特性的加权均方误差准则进行矢量量化,提高了量化增益。仿真结果表明,该方法实现简单,可以达到很好的压缩效果。     

Hierarchical finite-state vector quantization for image coding

The hierarchical finite-state vector quantization (HFSVQ) introduced in the paper is an improvement of the finite state vector quantization combined with hierarchical multirate image coding. Based on an understanding of the perception of human eye and the structural features of images, the HFSVQ technique employs different coding rates and different numbers of the predictive states for representative vector selection. The bit rate used to encode images is very low while the reconstructed images can still achieve a satisfactory perceptual quality     

Analog VLSI Circuits for Competitive Learning Networks

An investigation is made concerning implementations of competitive learning algorithms in analog VLSI circuits and systems. Analog and low power digital circuits for competitive learning are currently important for their applications in computationally-efficient speech and image compression by vector quantization, as required for example in portable multi-media terminals. A summary of competitive learning models is presented to indicate the type of VLSI computations required, and the effects of weight quantization are discussed. Analog circuit representations of computational primitives for learning and evaluation of distortion metrics are discussed. The present state of VLSI implementations of hard and soft competitive learning algorithms are described, as well as those for topological feature maps. Tolerance of learning algorithms to observed analog circuit properties is reported. New results are also presented from simulations of frequency-sensitive and soft competitive learning concerning sensitivity of these algorithms to precision in VLSI learning computations. Applications of these learning algorithms to unsupervised feature extraction and to vector quantization of speech and images are also described.     

Inverse error-diffusion using classified vector quantization

This correspondence extends and modifies classified vector quantization (CVQ) to solve the problem of inverse halftoning. The proposed process consists of two phases: the encoding phase and decoding phase. The encoding procedure needs a codebook for the encoder which transforms a halftoned image to a set of codeword-indices. The decoding process also requires a different codebook for the decoder which reconstructs a gray-scale image from a set of codeword-indices. Using CVQ, the reconstructed gray-scale image is stored in compressed form and no further compression may be required. This is different from the existing algorithms, which reconstructed a halftoned image in an uncompressed form. The bit rate of encoding a reconstructed image is about 0.51 b/pixel.     

Subband coding of images using asymmetrical filter banks

We address the problem of the choice of subband filters in the context of image coding. The ringing effects that occur in subband-based compression schemes are the major unpleasant distortions. A new set of two-band filter banks suitable for image coding applications is presented. The basic properties of these filters are linear phase, perfect reconstruction, asymmetric length, and maximum regularity. The better overall performances compared to the classical QMF subband filters are explained. The asymmetry of the filter lengths results in a better compaction of the energy, especially in the highpass subbands. Moreover, the quantization error is reduced due to the short lowpass synthesis filter. The undesirable ringing effect is considerably reduced due to the good step response of the synthesis lowpass filter. The proposed design takes into account the statistics of natural images and the effect of quantization errors in the reconstructed images, which explains the better coding performance.     

Fractal coding of subbands with an oriented partition

We propose a new image compression scheme based on fractal coding of the coefficients of a wavelet transform, in order to take into account the self-similarity observed in each subband. The original image is first decomposed into subbands containing information in different spatial directions and at different scales, using an orthogonal wavelet-generated filter bank. Subbands are encoded using local iterated function systems (LIFS), with range and domain blocks presenting horizontal or vertical directionalities. Their sizes are defined according to the correlation lengths and resolution of each subband. The edge degradation and the blocking effects encountered at low bit-rates using conventional LIFS algorithm are reduced with this approach. The computation complexity is also greatly decreased by a 12:1 factor in comparison to fractal coding of the full resolution image. The proposed method is applied to standard test images. The comparison with other fractal coding approaches and with JPEG shows an important increase in terms of PPSNR/bit-rate. Especially for images presenting a privileged directionality, the use of adaptive partitions results in about 3 dB improvement in PPSNR. We also discuss the distorsion versus rate improvement obtained on high-frequency subbands when fractal coding instead of pyramidal vector quantization is used. Our approach achieves a real gain in PPSNR for low bit-rates between 0.3 and 1.2 bpp.