Ihsan Ul Haq

Three-Class Mammogram Classification Based on Descriptive CNN Features

In this paper, a novel classification technique for large data set of mammograms using a deep learning method is proposed. The proposed model targets a three-class classification study (normal, malignant, and benign cases). In our model we have presented two methods, namely, convolutional neural network-discrete wavelet (CNN-DW) and convolutional neural network-curvelet transform (CNN-CT). An augmented data set is generated by using mammogram patches. To enhance the contrast of mammogram images, the data set is filtered by contrast limited adaptive histogram equalization (CLAHE). In the CNN-DW method, enhanced mammogram images are decomposed as its four subbands by means of two-dimensional discrete wavelet transform (2D-DWT), while in the second method discrete curvelet transform (DCT) is used. In both methods, dense scale invariant feature (DSIFT) for all subbands is extracted. Input data matrix containing these subband features of all the mammogram patches is created that is processed as input to convolutional neural network (CNN). Softmax layer and support vector machine (SVM) layer are used to train CNN for classification. Proposed methods have been compared with existing methods in terms of accuracy rate, error rate, and various validation assessment measures. CNN-DW and CNN-CT have achieved accuracy rate of 81.83% and 83.74%, respectively. Simulation results clearly validate the significance and impact of our proposed model as compared to other well-known existing techniques.

Referral system for hard exudates in eye fundus

Hard exudates are one of the most common anomalies/artifacts found in the eye fundus of patients suffering from diabetic retinopathy. These exudates are the major cause of loss of sight or blindness in people having diabetic retinopathy. Diagnosis of hard exudates requires considerable time and effort of an ophthalmologist. The ophthalmologists have become overloaded, so that there is a need for an automated diagnostic/referral system. In this paper a referral system for the hard exudates in the eye-fundus images has been presented. The proposed referral system works by combining different techniques like Scale Invariant Feature Transform (SIFT), K-means Clustering, Visual Dictionaries and Support Vector Machine (SVM). The system was also tested with Back Propagation Neural Network as a classifier. To test the performance of the system four fundus image databases were used. One publicly available image database was used to compare the performance of the system to the existing systems. To test the general performance of the system when the images are taken under different conditions and come from different sources, three other fundus image databases were mixed. The evaluation of the system was also performed on different sizes of the visual dictionaries. When using only one fundus image database the area under the curve (AUC) of maximum 0.9702 (97.02%) was achieved with accuracy of 95.02%. In case of mixed image databases an AUC of 0.9349 (93.49%) was recorded having accuracy of 87.23%. The results were compared to the existing systems and were found better/comparable.

Hyperspectral Unmixing Using Statistics of Q Function

Proposed technique of hyperspectral unmixing is apparent to implement and compute the results in a very fast and efficient manner. To reducing the computational complexity and to estimation of hyperspectral data we adopted a statistical method of median absolute deviation about median. Number of end-members is enumerating by self iterative subspace projection method which depends on Pearson correlation. The mixing matrix is inferred by using Q function projections. A set of tests with real hyperspectral data evaluates the performance and illustrates the effectiveness of the proposed method. For the evaluation of proposed method, the results are compared with the results of vertex component analysis. The experimental results show the effectiveness of proposed method on hyperspectral unmixing. targets Alunite, Buddingtonite, Calcite, Kaolinite, and Muscovite are detected well and have high spectral similarities. Hyperspectral remote sensing is used in a large array of real life applications e.g. Surveillance, Mineralogy, Physics, and Agriculture. The complete work is prepared by using MATLAB.

Wavelet Based De-noising Using Logarithmic Shrinkage Function

Noise in signals and images can be removed through different de-noising techniques such as mean filtering, median filtering, total variation and filtered variation techniques etc. Wavelet based de-noising is one of the major techniques used for noise removal. In the first part of our work, wavelet transform based logarithmic shrinkage technique is used for de-noising of images, corrupted by noise (during under-sampling in the frequency domain). The logarithmic shrinkage technique is applied to under-sampled Shepp–Logan Phantom image. Experimental results show that the logarithmic shrinkage technique is 7–10% better in PSNR values than the existing classical techniques. In the second part of our work we de-noise the noisy, under-sampled phantom image, having salt and pepper, Gaussian, speckle and Poisson noises through the four thresholding techniques and compute their correlations with the original image. They give the correlation values close to the noisy image. By applying median or wiener filter in parallel with the thresholding techniques, we get 30–35% better results than only applying the thresholding techniques individually. So, in the second part we recover and de-noise the sparse under-sampled images by the combination of shrinkage functions and median filtering or wiener filtering.

Automated System for Referral of Cotton-Wool Spots.

BACKGROUND Cotton-wool spots also referred as soft exudates are the early signs of complications in the eye fundus of the patients suffering from diabetic retinopathy. Early detection of exudates helps in the diagnosis of the disease and provides better medical attention. METHODS In this paper, an automated system for the detection of soft exudates has been suggested. The system has been developed by the combination of different techniques like Scale Invariant Feature Transform (SIFT), Visual Dictionaries, K-means clustering and Support Vector Machine (SVM). RESULTS The performance of the system is evaluated on a publically available dataset and AUC of 94.59% is achieved with the highest accuracy obtained is 94.59%. The experiments are also performed after mixing three datasets and AUC of 92.61% is observed with 91.94% accuracy. CONCLUSION The proposed system is easy to implement and can be used by medical experts both online and offline for referral of Cotton-wool spots in large populations. The system shows promising performance.

Hyperspectral Blind Unmixing and Multiple Target Detection Using Linear Mixture Model

In this paper a blind source separation technique Joint Approximate Diagonalization of Eigen-matrices (JADE) is investigated to unmixing and multiple target detection for hyperspectral imagery data. Our targeted minerals are Alunite, Buddingtonite, Calcite and Kaolinite in ‘Cuprite’ scene data that has been widely used for research experiments in hyperspectral imagery. A comparative study is conducted to show the effectiveness of the JADE with Vertex Component Analysis. The results are evaluated with both full and reduced bands.

In Vitro Biological Screening of Hartmannia rosea Extracts

The present study is focused on the assessment of the medicinal therapeutic potential extracts of H. rosea to investigate their pharmacological implications based upon science proofs. The antioxidant activity of fraction of H. rosea, namely, n-hexane (HR-1), ethyl acetate (HR-2), chloroform (HR-3), and n-butanol (HR-4), was performed by using the DPPH radical scavenging method. The cytotoxicity and enzyme inhibition assessment were also performed. All the extracts showed significant antioxidant, antibacterial, and protein kinase inhibition but none of the extracts exhibited α-amylase inhibition activity. The chloroform extract HR-3 may block a kinase receptor from binding to ATP; the lead molecule will be isolated, which may stop cancerous cell growth and demotion of cell division. It is predicted that ethyl acetate, chloroform, and n-butanol extracts of H. rosea contain polyphenolics, flavonoids, and alkaloids that are biologically effective candidates exhibiting significant cytotoxicity, antioxidant, and antimicrobial activities. They may control oxidative damage in the body tissues and act as potential antidiabetic and anticancer agents. These studies will also be helpful for future drug designing and drug development research.

Kernel fractional affine projection algorithm

AbstractThis paper extends the kernel affine projection algorithm to a rich, flexible and cohesive taxonomy of fractional signal processing approach. The formulation of the algorithm is established on the inclusion of Riemann–Liouville fractional derivative to gradient-based stochastic Newton recursive method to minimize the cost function of the kernel affine projection algorithm. This approach extends the idea of fractional signal processing in reproducing kernel Hilbert space. The proposed algorithm is applied to the prediction of chaotic Lorenz time series and nonlinear channel equalization. Also the performance is validated in comparison with the least mean square algorithm, kernel least mean square algorithm, affine projection algorithm and kernel affine projection algorithm.