Hammad Qureshi

Adaptive Discriminant Wavelet Packet Transform and Local Binary Patterns for Meningioma Subtype Classification

The inherent complexity and non-homogeneity of texture makes classification in medical image analysis a challenging task. In this paper, we propose a combined approach for meningioma subtype classification using subband texture (macro) features and micro-texture features. These are captured using the Adaptive Wavelet Packet Transform (ADWPT) and Local Binary Patterns (LBPs), respectively. These two different textural features are combined together and used for classification. The effect of various dimensionality reduction techniques on classification performance is also investigated. We show that high classification accuracies can be achieved using ADWPT. Although LBP features do not provide higher overall classification accuracies than ADWPT, it manages to provide higher accuracy for a meningioma subtype that is difficult to classify otherwise.

Hyperspectral Colon Tissue Classification using Morphological Analysis

Diagnosis and cure of colon cancer can be improved by efficiently classifying the colon tissue cells into normal and malignant classes. This paper presents the classification of hyperspectral colon tissue cells using morphological analysis of gland nuclei cells. The application of hyperspectral imaging technique in medical image analysis is a new domain for researchers. The main advantage in using hyperspectral imaging is the increased spectral resolution and detailed subpixel information. Biopsy slides with several microdots, where each microdot is from a distinct patient, are illuminated with a tuned light source and magnification is performed up to 400times. The proposed classification algorithm combines the hyperspectral imaging technique with linear discriminant analysis. Dimensionality reduction and cellular segmentation is achieved by independent component analysis (ICA) and k-means clustering. Morphological features, which describe the shape, orientation and other geometrical attributes, are next to be extracted. For classification, LDA is employed to discriminate tissue cells into normal and malignant classes. Implementation of LDA is simpler than other approaches; it saves the computational cost, while maintaining the performance. The algorithm is tested on a number of samples and its applicability is demonstrated with the help of measures such as classification accuracy rate and the area under the convex hull of ROC curves

Handwritten digit recognition through wavelet decomposition and wavelet packet decomposition

Handwritten digit recognition is a significant and established problem in computer vision and pattern recognition and a lot of research work has already been carried out in this area. In this paper a new technique for handwritten digit recognition is proposed. As the handwritten digits are not of the same size, thickness, style, position and orientation therefore different challenges have to be faced to resolve the problem of handwritten digit recognition. The uniqueness and variety in the writing styles of different people also influence the pattern and appearance of the digits. Handwritten digit recognition is the method of recognizing and classifying handwritten digits. It has wide application such as automatic processing of bank cheques, postal addresses and tax forms etc. In this paper, we present a wavelets analysis based technique for feature extraction. The task of classification is handled using KNN and SVM classifier. An overall high recognition rate of 97.04 is achieved on the test data set. The proposed scheme is tested on the well known MNIST data set.

Monitoring Disease Outbreak through Geographical Representation in Rural Areas

Geographical analysis and tracking of the spread of epidemics and other diseases is an important issue and is of gre¬¬¬at concern to healthcare professionals all over the world. Information technology has played a vital role in building tools for spatial analysis of spread of diseases and hence has made it possible to track epidemics and early adoption of preventive measures. In this paper, we present a model eSystem for spatial plotting of patients for a disease using our Jaroka Tele-Healthcare System (JTHS) and Google Maps. We use the existing Information and Communication Technology (ICT) infrastructure to provide healthcare to people in remote and rural areas of Pakistan. Lady and Community Health Workers (LCHWs) use simple mobile phones to register patients and report symptoms using the Short Messaging Service (SMS) with the JTHS. Further correspondence for prescription and follow-up diagnosis of the registered patients can also be carried out using the system. This is especially important for remote areas where medical facilities are either rare or completely unavailable. JTHS enables doctors, medical experts and health officials to view the geospatial location of more than twenty thousand patients on the click of a button. This model has been successfully implemented in the rural areas of Mardan, KPK Pakistan and is now being replicated in other parts of the country. The system is unique in the sense that it uses mobile phones based crowd-sourcing to register patients and enable health professionals to track spread of diseases. We also show how population migration affected disease incidence in rural Mardan and how JTHS was able to geographically display the spread of disease.

Wavelet transforms and wavelet packets in bio-medical engineering

Wavelets based analysis has become very popular in a wide area of applications. It has found application in areas ranging from image and video compression to texture analysis and image recognition. Using wavelet analysis, multi-resolution analysis is performed which means that an image is decomposed in to multiple time frequency or scale space resolutions. This leads to decomposition of an image in to various representations allowing for effective analysis of the data. In this tutorial, we would should how wavelets-based analysis may be used to acquire interesting features for analysis and segmentation of medical images for computer assisted diagnosis. Medical image analysis for computer assisted diagnosis falls under the domain of biomedical engineering and has become a very important area of research as it can help in reduction of errors made in diagnosis by physicians and doctors. The aims of the tutorial are as under: Develop detailed understanding of wavelet analysis using wavelet transform and wavelet packets; Understand how multi-resolution analysis may be performed using wavelets; Apply the technique to a wide variety of problems from the domain of medicine for computer assisted diagnosis.

Parameter evaluation for virtual Laparoscopic simulation

Virtual Reality based surgical simulators have become quite common for training of surgeons for different surgical skills. Simulators have been widely used particularly in minimal invasive surgery. In this paper we find parameters that would be required to create a real time working simulation for exercises given in the Fundamentals of Laparoscopic Surgery curriculum. We use peg transfer exercise as our example in this work and create simulations for parameter analysis using SOFA, an open source surgical framework [1]. The parameters we choose are generic and can be used to create other more complex simulations like cholecystectomy [2] (gall bladder removal) and appendectomy (appendix removal). We show the implementation of these parameters and their behavior in a virtual reality surgical simulation. This work can be used by researchers and developers to choose the right parameters in the context of the simulation they are developing. It also shows the cost and behavior of achieving good visualization (frames per second), physical characteristics and a realistic behavioral model to be used in simulations for training purposes.

INAV: Minimizing delay induced by DCF control packet losses in IEEE 802.11 to optimize throughput

IEEE 802.11 uses Distributed Coordination Function (DCF) as default channel access method which is based on Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) where all the nodes compete for the channel access. Based on this mechanism a node cannot transmit until the medium is free. Collision can occur if two or more stations within each others transmission range transmit simultaneously. To overcome this problem several solutions have been proposed, for instance Request To Send/Clear To Send (RTS/CTS) mechanism, through which the channel is reserved for a specified time period known as Network Allocation Vector (NAV). Although this mechanism provide all the nodes to share the medium fairly but it is observed that this mechanism can lead to bandwidth underutilization, especially in networks with higher packet delivery failure rates. In this paper, we propose Intimated Network Allocation Vector (INAV) - a modified version of existing NAV mechanism - which considerably reduce bandwidth underutilization in such networks. This modification works by exploiting the backing-off feature of neighbors of the communicating nodes. The proposed technique recovers more than 93% of idle waiting time for neighbors of the nodes that fail to complete the four way handshake of RTS/CTS.

Association of pre-pregnancy weight and weight gain with perinatal mortality

Reducing infant mortality is one of the primary Millennium Development Goals 2015. A lot of effort has been made to reduce infant mortality but it remains high in most of the developing countries and the underdeveloped world. Perinatal Mortality is a cause of great emotional pain and social unrest. The main cause of pregnancy failure in the developed world is obesity but in the under-developed world the main cause remains malnutrition. However, their are a mix of factors that affect pregnancy failure in the developing countries. Pakistan has a very high infant mortality rate which stands at 78 deaths per 1000 births. The reasons for this are many including lack of proper healthcare. This is because of a severe shortage of healthcare professionals and specialists in Pakistan. The gap in healthcare may be overcome by leveraging IT to provide automated healthcare. In this paper, we show how machine learning may be used to predict perinatal failure. We examine the relationship between pre-pregnancy weight, weight gain during pregnancy and the body mass index (BMI) to investigate how they relate to foetal failure. We employ the K Nearest Neighbor (K-NN) technique to automatically differentiate between successful and failed pregnancies. Our method is able to predict the the outcome of a pregnancy with about 95% accuracy.