Department of Computer Science

Permanent URI for this communityhttps://repository.cuilahore.edu.pk/handle/123456789/16

Browse

Author: search.filters.author.Dr. Usama Ijaz BajwaSubject: search.filters.subject.Computer Science

Search Results

Now showing 1 - 6 of 6
  • No Thumbnail Available
    Item
    Classification of MGMT Promoter Methylation Status in Brain Tumor MR Images using Deep Learning
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2023) Muhammad Sohaib Iqbal; FA21-RCS-005; LHR TP 8470; Dr. Usama Ijaz Bajwa
    Glioblastoma is the most dangerous brain tumor type (worst prognosis), having the least survival rate of less than a year. O6-methylguanine-DNA methyltransferase (MGMT) promoter status “methylation” adds a favorable prognosis and is a reliable indication of cancer's chemosensitivity. Therefore, determining the MGMT status of a brain tumor patient is essential for treatment planning. Surgical tissue sample removal is required to conduct genetic testing for tumors. The genetic characterization of the tumors may then take a few weeks to establish after that. Non-invasive diagnosis using MR images plays a significant role in the early detection of the deadly disease by using Magnetic Resonance Imaging (MRI) to detect the tumorous regions using deep learning techniques. The main requirement for detecting MGMT status is obtaining the sub-regions of brain tumor from MRIs. Previous studies have worked on Brain tumor segmentation and MGMT classification separately, few of the studies have worked on segmentation and classification together but have used selective modalities for MGMT Prediction. Since each modality has its own significance and MGMT promoter status lies in the tumorous region, this study proposed a pipeline of segmentation and classification models using stacked multimodalities of MRI scans of the benchmark dataset BraTS2021. The pipeline is divided into two phases. The first phase uses a stack of multiple MRI modalities to segment the brain tumor into sub-region using a 3D Residual U-Net Architecture, and the second phase uses the segmentation model's output (stacked multimodalities tumor voxel) to determine the MGMT promoter status using a 3D ResNet 10 Classifier. With the help of pipeline, classification model received precise information about the tumorous regions which helped in efficient prediction of the MGMT status present in the brain tumor. The proposed pipeline has therapeutic value because it can assist radiologists in diagnosing brain tumors more efficiently and precisely by reducing the subjectivity and variability of human interpretation. The segmentation and classification pipeline for brain tumors x can potentially increase the consistency and objectivity of diagnoses. Furthermore, it can also assist radiologists in predicting the MGMT status without using surgical equipment and can help in treatment decision making. The segmentation phase of the proposed pipeline yielded promising results with average dice scores of 0.81, 0.84, and 0.80 for the tumor core (TC), whole tumor (WT), and enhancing tumor (ET) on validation. The classification phase achieved ROC-AUC score of 0.66 on validation.
  • No Thumbnail Available
    Item
    Enhancing Crime Classification in Surveillance Videos for Real Time Monitoring
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2021) Muhammad Salman Ghauri; SP19-RCS-027; LHR TP 7590; Dr. Usama Ijaz Bajwa
    Anomalies can be detected with the help of patterns and events that differ from the normal flow of events. The paradigms of surveillance may relate to abuse, fights, road accidents and snatchings, etc. In real-world surveillance, finding unusual events in these massive video streams is a difficult endeavor, since they often occur inconsistently. However, deep learning-based anomaly detection helps reduce human labor and its decision-making ability can be compared to that of humans, thus ensuring the safety of the public. In the majority of reported studies, anomalies are detected from surveillance videos based on binary classification. The reported approaches did not cover other anomalous events from surveillance videos, including abuse, fights, vehicle accidents, shootings, stealing, vandalism, and robberies. This paper proposes an intelligent anomaly detection framework based on deep features that can operate more efficiently in surveillance networks. In the proposed framework, spatial-temporal features are first extracted from a series of frames by passing them through a CNN model that has been pretrained. Analyzing the frames in a sequence can be beneficial in detecting anomalous events. Once the deep features have been extracted, the data is then passed to the Long Short-Term Memory (LSTM) model. The model can accurately classify ongoing anomalies/normal events in complex surveillance scenes of smart cities. A dataset from the University of Central Florida (UCF) Crime video dataset is used to perform extensive experiments on anomaly detection. We report an increase in data accuracy of 47.83% over state-of-the-art methods for UCF-Crime datasets.
  • No Thumbnail Available
    Item
    Transfer Learning based Deep Learning Approach for Brain Tumor Segmentation from MR Images
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2021) Rehan Raza; SP20-RCS-018; LHR TP 7594; Dr. Usama Ijaz Bajwa
    Glioma is the most prevalent and dangerous type of brain tumor which can be life-threatening when its grade is high. The early detection of these tumors can improve and save the life of the patients. The automatic segmentation of brain tumors from magnetic resonance imaging (MRI) plays a vital role in treatment planning and timely diagnosis. Automatic segmentation is a challenging task due to the massive amount of information provided by MRI and the variation in the location, type, and size of the tumor. Therefore, a reliable and authentic method to segment the tumorous region from healthy tissues accurately is an open challenge in the field of deep learning-based medical image analysis. This thesis presents an end-to-end framework for automatic 3D Brain Tumor Segmentation (BTS). The proposed model is a hybrid of the deep residual network and U-Net model (dResU-Net). The residual network is used as an encoder in the proposed architecture with the decoder of the U-Net model to handle the issue of vanishing gradient. The proposed model is designed to take benefit from low and high features. In addition, shortcut connections are employed in residual convolutional blocks and skip connections between residual and convolutional blocks are utilized in the proposed architecture to accelerate the training process. The proposed architecture achieved promising results with the average dice score of 0.8357, 0.8660, and 0.8004 for the tumor core (TC), whole tumor (WT), and enhancing tumor (ET), respectively, on BraTS 2020 dataset. Furthermore, the proposed dResU-Net model is also compared with modified 3D U-Net and pre-trained ResU-Net34 and ResU Net50 to check the performance of the proposed model over the pre-trained models. To demonstrate the robustness of the proposed model in real-world clinical settings, validation of the trained model on an external cohort is performed on randomly selected 50 patients of the BraTS 2021 benchmark dataset. The achieved dice scores on external cohorts are 0.8400, 0.8601, and 0.8221 for TC, WT, and ET, respectively. The comparison of results of the proposed technique with the state-of-the-art techniques indicates that dResU- x Net can significantly improve the segmentation performance of brain tumor sub-regions.
  • No Thumbnail Available
    Item
    Deception Detection using Facial Action Coding System in Videos
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2020) Hammad-ud-din Ahmed; SP17-RCS-015; LHR TP 5974; Dr. Usama Ijaz Bajwa
    Humans lie to each other on a regular basis in one form or another. To lie to someone is to share altered facts. Facts are important for decision making in every situation. Without unaltered facts, potentially harmful decisions can be made which can alter someone’s life in ways that may not want. This is why it is important to catch misinformation before any harm can be caused by them. Deception detection in videos has gained traction in recent times for its various real-life applications. When a person lies, they show facial expressions, also known as microexpressions so fast that they can be ignored by the untrained eye. Microexpressions can be used as a basis for creating a deception detection system. Facial Action Coding System is utilized as a way to encode and extract data from facial muscle movement during truthful and deceptive confessions. This data is used to train a deep learning model that utilizes long short-term memory (LSTM) to train and create the system. The real-life trial dataset is used to train and test the system provided one of the best facial only approaches to deception detection. Cross-data validation is also tested using the Real-life trial dataset, the Silesian Deception Database, and the Bag-of-lies Deception Database. Cross-data validation is something that has not yet been attempted by anyone else for a deception detection system which provides unique insight for the field. The results show that adding different datasets to train a neural network for the sake of creating a deception detection system worsens the accuracy of the system but the reasoning behind the poorer results actually encourages attempts at creating better datasets
  • No Thumbnail Available
    Item
    Automated Detection and Classification of Brain Tumor from MRI Images using Machine Learning Methods
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2018) Ghulam Gilanie; FA14-PCS-007; LHR TP 5805; Dr. Usama Ijaz Bajwa
    The focus of this thesis is to report an automated, efficient, and robust method of brain tumor detection and classification from Magnetic Resonance Images (MRI) images. Clinically, it is a challenging issue faced by the researchers working in this domain. In routine health care units, Magnetic Resonance (MR) scanners are being used to generate a massive number of brain slices, underlying the anatomical details. Pathological assessment from this medical data is being carried out manually by the radiologists or neuro oncologists. Due to huge volume of brain anatomical data produced by MRI scanners, it is almost impossible to manually analyze every slice. Conclusively, if automated protocols are executed for auto-interpretation; not only the radiologist will be assisted but also a better pathological assessment process would be expected. Several methods have been suggested to address this problem, but still, accuracy, robustness and optimization is still an open issue to address. The development of such automated procedures is difficult due to complex organization of brain cells, several types of tumor, difference in medical traits of a specific ethnicity and many more factors. To achieve the target, research has been started from reviewing the most popular and prominent state-of-the-art methods. Based upon the reviewed literature, automated brain tumor detection and classification techniques have been reported with high computational cost, low classification rates, detection and classification of only one or a few of brain tumor types, lack of robustness, etc. Therefore, step wise research and experiments based upon empirical scientific methodology have been performed in order to achieve the objectives of brain tumor classification. In the first step, a research activity has been performed to report a colorization method with the aims to enhance the visualization, cell characterization and interpretation of brain cells. The high dimensional brain data scanned through MRI embodied in gray scale, if converted, represented, mapped and/or visualized in colored versions, irrefutably, more definitive and more accurate the pathological assessment process will be. Several methods have been reported to represent brain MRI data in color with high computational xii complexity. In this research activity, an efficient method of colorization using frequencies from visible range of color spectrum, has been proposed to embody the variations and sensitivity of the brain MRI images. The experiments have been performed on a locally developed dataset. Side by side visual comparison based on multiple MRI sequences of identical subjects by domain experts have proved the adequate success and fruitfulness of the story. The reported method of colorization as a protocol has also been deployed in Department of Radiology and Diagnostic Images, Bahawal Victoria, Hospital, Bahawalpur (BVHB), Pakistan. Radiologists are using this tool for visual interpretation and monitoring of the patients for their assessment and clinical decision making. In second step, an automated approach using Gabor filter and Support Vector Machines (SVMs), for the classification of brain MRI slices as normal or abnormal has been reported. Accuracy, sensitivity, specificity and AUC-value have been used as standard quantitative measures to evaluate the proposed algorithm. To the best of our knowledge, this is the first study in which experiments have been performed on The Whole Brain Atlas - Harvard Medical School (HMS) dataset, achieving an accuracy of 97.5%, sensitivity of 99%, specificity of 92% and AUC-value as 0.99. To test the robustness against medical traits based on ethnicity and to achieve optimization, a locally developed dataset has also been used for experiments and remarkable results with accuracy (96.5%), sensitivity (98%), specificity (92%) and AUC-value (0.97) were achieved. Comparison with state-of-the art methods proved the overall efficacy of the proposed method. In third step of the thesis, a method has been proposed to classify brain MRI image into brain related disease groups and further tumor types. The proposed method employed Gabor texture followed by a set of more distinguished statistical features. These features are then used by SVM to classify the brain disorder. K-fold strategy has been adapted for cross validation of the results to enhance generalization of SVM. Experiments have been performed to classify brain MRI images as normal or belonging to either of the common diseases, such as cerebrovascular, degenerative, inflammatory, and neoplastic. Neoplastic disease is further classified into glioma, meningioma, metastatic adenocarcinoma, metastatic bronchogenic carcinoma, or sarcoma. Standard quantitative evaluation measures, i.e., accuracy, specificity, sensitivity, and AUC-value have been used to test xiii performance of the developed system. The proposed system has been trained on complete dataset of HMS, so the trained model has the ability to deal with a wide range of brain abnormalities. Further, to achieve robustness, a locally developed dataset has also been used for experiments. Remarkable results on different orientations, sequences of both of these datasets as per accuracy (up-to 99.6%), sensitivity (up-to 100%), specificity (up-to 100%), precision (up-to 100%) and AUC-value (up-to 1.0) have been achieved. The proposed method classifies the brain MRI slices into defined abnormality groups. It can also classify the abnormal slices into tumorous or non-tumorous one. The major achievement of the developed system is its auto classification of tumorous slices into the slices having primary tumor or secondary tumor and their further types, which possibly could not be determined without biopsy. In fourth step of the thesis, results achieved through the proposed method of brain tumor classification have been validated on cross data set. The drive of this research activity is to verify the robustness of the reported approach. For this, the model has been trained completely on one data set, while tested completely on another one. A benchmarked dataset HMS and a locally developed dataset BVHB dataset has been used for this purpose. To ensure its robustness, complete HMS dataset was used to train the model and BVHB was used to test the trained model and vice versa. Standard evaluation measures, i.e., accuracy, specificity, sensitivity, precision and AUC-value have been used to evaluate the system. It has been established that the proposed method deals with multiformity and variability of brain MRI data. Overall, suppositions regarding robustness of the proposed method were attained with maximum measures as per accuracy as 92%, specificity as 92%, sensitivity as 93%, precision as 92%, and AUC-value as 0.93. The overall results achieved through the proposed method, manifests that it is robust, efficient and reliable. It has been trained on a large volume of multi-orientations, multi-sequences belonging to multi-datasets to deal with multiformity and to face variabilit
  • No Thumbnail Available
    Item
    Automated Detection of Early Pulmonary Nodule in Computed Tomography Images
    (Library Information Services, COMSATS University Islamabad, Lahore Campus, 2019) Ahmad Usama Tariq; FA16-RCS-012; LHR TP 5777; Dr. Usama Ijaz Bajwa
    Classification of lung cancer in CT scans majorly have two steps, detect all suspicious lesions also known as pulmonary nodules and calculate the malignancy. Currently, a lot of studies are about nodules detection, but some are about the evaluation of nodule malignancy. Since the presence of nodule does not unquestionably define the presence lung cancer and the morphology of nodule has a complex association with malignant growth, the diagnosis of lung cancer requests cautious examinations on each suspicious nodule and integrateed information every nodule. We propose a 3D CNN CAD system to solve this problem. The system consists of two modules a 3D CNN for nodule detec tion, which outputs all suspicious nodules for a subject and second module train on XGBoost classifier with selective data to acquire the probability of lung malignancy for the subject