Performance analysis of support vector machines classifiers in breast cancer mammography recognition

Support vector machine (SVM) is a supervised machine learning approach that was recognized as a statistical learning apotheosis for the small-sample database. SVM has shown its excellent learning and generalization ability and has been extensively employed in many areas. This paper presents a performance analysis of six types of SVMs for the diagnosis of the classical Wisconsin breast cancer problem from a statistical point of view. The classification performance of standard SVM (St-SVM) is analyzed and compared with those of the other modified classifiers such as proximal support vector machine (PSVM) classifiers, Lagrangian support vector machines (LSVM), finite Newton method for Lagrangian support vector machine (NSVM), Linear programming support vector machines (LPSVM), and smooth support vector machine (SSVM). The experimental results reveal that these SVM classifiers achieve very fast, simple, and efficient breast cancer diagnosis. The training results indicated that LSVM has the lowest accuracy of 95.6107 %, while St-SVM performed better than other methods for all performance indices (accuracy = 97.71 %) and is closely followed by LPSVM (accuracy = 97.3282). However, in the validation phase, the overall accuracies of LPSVM achieved 97.1429 %, which was superior to LSVM (95.4286 %), SSVM (96.5714 %), PSVM (96 %), NSVM (96.5714 %), and St-SVM (94.86 %). Value of ROC and MCC for LPSVM achieved 0.9938 and 0.9369, respectively, which outperformed other classifiers. The results strongly suggest that LPSVM can aid in the diagnosis of breast cancer.     

Dynamic watermarking scheme for quantum images based on Hadamard transform

In this paper, a novel watermarking scheme for quantum images based on Hadamard transform is proposed. In the new scheme, a unitary transform controlled by a classical binary key is implemented on quantum image. Then, we utilize a dynamic vector, instead of a fixed parameter as in other previous schemes, to control the embedding process. The dynamic embedding vector is decided by both the carrier quantum image and the watermark image, which is only known by the authorized owner. The proposed scheme is analyzed from visual quality, computational complexity, and payload capacity. Analysis and results show that the proposed scheme has better visual quality under a higher embedding capacity and lower complexity compared with other schemes proposed recently.     

Piezoresistivity in Polymer-Ceramic Composites

Composite materials composed of randomly dispersed semiconducting ceramic particles in an insulating polymer matrix show a pronounced change in resistivity with pressure. Different amounts of iron oxide (Fe₃O₄) powder and antimony-doped tin oxide (SnO₂:Sb) powder were dispersed in an epoxy polymer matrix to form pressure-sensitive composites. In each family of materials, an insulator-to-semiconductor transition is observed in agreement with percolation theory. Composites within a certain range of filler content showed substantial piezoresistive effect under both uniaxial and hydrostatic pressure in which sensitivity is controlled by the choice of filler material and the volume fraction. The effect of temperature on the piezoresistance effect was also examined. Piezoresistors made from Fe₃O₄ composites showed larger temperature changes than those filled with Sb-doped SnO₂.     

Variations in fat content and lipid class composition in ten different mango varieties

The kernels of 10 different mango varieties were extracted. The physico-chemical characteristics and lipid class composition of fats were studied. The fat content of mango kernels grown under the soil and climatic conditions of Bangladesh varied from 7.1% to 10%, depending on the variety. The total lipid extracts were fractionated into lipid classes by a combination of column and thin layer chromatography (TLC). The hydrocarbon and sterol esters varied from 0.3% to 0.7%, triglycerides from 55.6% to 91.5%, partial glycerides from 2.3% to 4% and free sterol from 0.3% to 0.6%. Free fatty acids amounted to 3.0–37% as oleic; glycolipids were 0.6–1.2% and phospholipids 0.11–0.8%. The fatty acid composition of triglyceride (TG) fractions was analyzed by gas liquid chromatography (GLC). Palmitic acid varied from 7.9 molar % to 10.0 molar %, stearic from 38.2% to 40.2%, oleic from 41.1% to 43.8%, linoleic from 6.0% to 7.6%, linolenic from 0.6% to 1.0% and arachidic acid from 1.7% to 2.6%. TLC revealed the presence of lyso-phosphatidylcholine, phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine and phosphatidic acid in the phospholipid fraction.     

Linear principal transformation: toward locating features in N-dimensional image space

Projection Functions have been widely used for facial feature extraction and optical/handwritten character recognition due to their simplicity and efficiency. Because these transformations are not one-to-one, they may result in mapping distinct points into one point, and consequently losing detailed information. Here, we solve this problem by defining an N-dimensional space to represent a single image. Then, we propose a one-to-one transformation in this new image space. The proposed method, which we referred to as Linear Principal Transformation (LPT), utilizes Eigen analysis to extract the vector with the highest Eigenvalue. Afterwards, extrema in this vector were analyzed to extract the features of interest. In order to evaluate the proposed method, we performed two sets of experiments on facial feature extraction and optical character recognition in three different data sets. The results show that the proposed algorithm outperforms the observed algorithms in the paper and achieves accuracy from 1.4 % up to 14 %, while it has a comparable time complexity and efficiency.     

An image composition algorithm for handling global visual effects

Image composition is widely used in television and film industry to create synthetic visual effects. It requires seamless integration of different parts of two or more images into a single image. Existing image composition techniques only change the local contents of the resulting image while in many cases local changes may also require some global effects as well. For example, if the image of sun from one image is transferred to another image, the global brightness pattern should also be transferred. Unfortunately existing techniques cannot handle global effects of local content manipulations. This paper describes a novel image composition technique which captures global effects associated with a specific local content from one image and incorporates in the second image. In our proposed technique, all images are transformed to the frequency domain. The composite image is created in frequency domain by mixing different frequencies from multiple images and then transformed back to the spatial domain. We have experimented the proposed technique to shift the image of sun along with its global brightness pattern, the global effects of rain and also for transferring global texture pattern from one image to the other. In most of the cases the results produced by our algorithm appear far close to real images than state of the art existing image composition techniques.     

Predictive and content-aware load balancing algorithm for peer-service area based IPTV networks

Load balancing is a crucial factor in IPTV delivery networks. Load balancing aims at utilizing the resources efficiently, maximizing the throughput, and minimizing the request rejection rate. The peer-service area is the recent architecture for IPTV delivery networks that overcomes the flaws of the previous architectures. However, it still suffers from the load imbalance problem. This paper investigates the load imbalance problem, and tries to augment the peer-service area architecture to overcome this problem. To achieve the load balancing over the proposed architecture, we suggest a new load-balancing algorithm that considers both the expected and the current load of both contents and servers. The proposed load-balancing algorithm consists of two stages. The first stage is the contents replication according to their expected load, while the second stage is the content-aware request distribution. To test the effectiveness of the proposed algorithm, we have compared it with both the traditional Round Robin algorithm and Cho algorithm. The experimental results depict that the proposed algorithm outperforms the two other algorithms in terms of load balance, throughput, and request rejection rate.