fatimah khalid

SwinUNeLCsT: Global-local spatial representation learning with hybrid CNN-transformer for efficient tuberculosis lung cavity weakly supervised semantic segmentation

Journal of King Saud University. Computer and information sciences/Maǧalaẗ ǧamʼaẗ al-malīk Saud : ùlm al-ḥasib wa al-maʼlumat, Apr 1, 2024

Herbal Plant Image Classification using Transfer Learning and Fine-Tuning Deep Learning Model

Journal of Advanced Research in Applied Sciences and Engineering Technology, Dec 15, 2023

Medical Image Analysis of Lung Tumor Diagnosis Based on Generalized Fuzzy C-Means Clustering Algorithm

Journal of physics, May 1, 2023

In recent years, the rapid development of medical imaging technology has brought medical image an... more In recent years, the rapid development of medical imaging technology has brought medical image analysis into the era of big data. CT imaging technology is one of the most common imaging methods for disease screening. This paper introduces medical image analysis of lung tumor diagnosis based on Generalized fuzzy c - means clustering algorithm was proposed. Brief introduce lung CT imaging to diagnose lung tumors requires a huge workload, which is often accompanied by long-term image reading and subjective evaluation. The experimental data comes from 15 patient samples which include two groups of samples are selected as test objects. The result shows the positive rate is high, resulting in misdiagnosis. In addition, great success of deep learning technology in the field of computer vision has made it possible to realize computer-aided diagnosis and screening of lung cancer.

A novel two-stage Lung nodule detection in medical images based on deep learning

Journal of physics, May 1, 2023

In order to solve the problems of low detection rate and high false positive of nodules in lung C... more In order to solve the problems of low detection rate and high false positive of nodules in lung CT images by traditional computer-aided diagnosis systems. This paper introduces the real-time detection algorithm of medical images using the deep learning method. According to the three-dimensional nature of lung CT images, 3DFasterR-CNN is used to extract features to detect candidate nodules; then, the 3D convolutional neural network is used to remove false positive nodules. That saves the operation of filtering and difference map construction. The detection process mainly includes four steps: building a deep neural network, constructing suitable samples, training the network, and detecting. The results show that the algorithm has a good effect on the real-time change detection of medical images and can improve detection accuracy and efficiency. The average FROC value was 82.8%, compared with the traditional diagnosis and treatment. The recognition rate of the method has been significantly improved. The model can be used to assist doctors in diagnosing lung nodules and has a certain clinical application value.

SMARTFLORA Mobile Flower Recognition Application Using Machine Learning Tools

Hybrid Discrete Wavelet Transform and Histogram of Oriented Gradients for Feature Extraction and Classification of Breast Dynamic Thermogram Sequences

Breast cancer is responsible for the death of thousands of women around the world. A cancerous tu... more Breast cancer is responsible for the death of thousands of women around the world. A cancerous tumour increases temperature in the region near to the tumour, such heating is then transferred to the skin surface. Early breast cancer detection can save many lives and reduce the cost of treatment. There are many imaging techniques to do the screening such as thermography. To achieve an early detection, Computer-Aided Detection (CAD) systems are needed to classify masses in thermogram images. In this research, we propose a hybrid methodology based on Histogram of Oriented Gradients (HOG) and Discrete Wavelet Transform (DWT) to classify temporal-based thermogram images into either normal or cancerous. In this study we propose a technique to reduce the HOG coefficient vector and extract features from this vector. Two-dimensional Discrete Wavelet Transform (DWT) is applied on the original images to obtain HH (high-high) sub band. Then, HOG is applied using the HH sub band. Support Vector Machine (SVM) binary classifier is used to classify images to either normal or abnormal using the extracted features. The obtained results showed comparable results with state-of-the-art related research.

A new clustering approach for group detection in scene-independent dense crowds

Despite significant progress in crowd behaviour analysis over the past few years, most of today's... more Despite significant progress in crowd behaviour analysis over the past few years, most of today's state of the art algorithms focus on analysing individual behaviour in a specific-scene. Recently, the widespread availability of cameras and a growing need for public safety have shifted the attention of researchers in video surveillance from individual behavior analysis to group and crowd behavior analysis. However, dangerous and illegal behaviours are mostly occurred from groups of people. Group detection is the main process to separate people in crowded scene into different group based on their interactions. Results of group detection can further to apply in analyze group and crowd behaviour. This paper present a study of the group detection and propose a novel approach for clustering group of people in different crowded scenes based on trajectories. For the clustering of group of people we propose novel formula to compute the weights based on the distance, the occurrence, and the speed correlations of two people in a tracklet cluster to infer the people relationship in a tracklet clusters with Expectation Maximization (EM) in order to overcome occlusion in crowded scenes.

Download

Region-based surveillance video retrieval with effective object representation

The Imaging Science Journal, Nov 9, 2017

ABSTRACT Retrieving the most relevant video frames that contain the object specified in a given q... more ABSTRACT Retrieving the most relevant video frames that contain the object specified in a given query (query-by-region) remains a challenging task. Two common challenges of region-based retrieval approaches are to accurately extract or segment object(s) and select a proper matching strategy. This paper addresses these problems by proposing a retrieval approach that uses a new region-based matching technique equipped with an effective object representation method. In the first stage, the proposed approach selects the most informative instances of each object that appeared in the video by utilizing an adapted clustering algorithm over the extracted features. In the retrieval stage, the new matching technique returns the most relevant sequences of video by mapping a given region with those identified representative instances of objects based on their similarity scores. The proposed approach is evaluated on standard datasets and the results demonstrate a 31% improvement in the retrieval performance compared to other state-of-the-art methods.

Speeded up surveillance video indexing and retrieval using abstraction

Many researches have been conducted on video abstraction for quick viewing of video archives, how... more Many researches have been conducted on video abstraction for quick viewing of video archives, however there is a lack of approach that considers abstraction as a pre-processing stage in video analysis. This paper aims to investigate the efficiency of integrating video abstraction in surveillance video indexing and retrieval framework. The basic idea is to reduce the computational complexity and cost of overall processes by using the abstract version of the original video that excludes unnecessary and redundant information. The experimental results show a significant reduction of 87% in computational cost by using the abstract video rather than the original video in both indexing and retrieval processes.

Download

Hybrid Discrete Wavelet Transform and Histogram of Oriented Gradients for Feature Extraction and Classification of Breast Dynamic Thermogram Sequences

2021 Fifth International Conference on Information Retrieval and Knowledge Management (CAMP)

Breast cancer is responsible for the death of thousands of women around the world. A cancerous tu... more Breast cancer is responsible for the death of thousands of women around the world. A cancerous tumour increases temperature in the region near to the tumour, such heating is then transferred to the skin surface. Early breast cancer detection can save many lives and reduce the cost of treatment. There are many imaging techniques to do the screening such as thermography. To achieve an early detection, Computer-Aided Detection (CAD) systems are needed to classify masses in thermogram images. In this research, we propose a hybrid methodology based on Histogram of Oriented Gradients (HOG) and Discrete Wavelet Transform (DWT) to classify temporal-based thermogram images into either normal or cancerous. In this study we propose a technique to reduce the HOG coefficient vector and extract features from this vector. Two-dimensional Discrete Wavelet Transform (DWT) is applied on the original images to obtain HH (high-high) sub band. Then, HOG is applied using the HH sub band. Support Vector Machine (SVM) binary classifier is used to classify images to either normal or abnormal using the extracted features. The obtained results showed comparable results with state-of-the-art related research.

Phishing Image Spam Classification Research Trends: Survey and Open Issues

International Journal of Advanced Computer Science and Applications, 2020

A phishing email is an attack that focused completely on people to circumvent existing traditiona... more A phishing email is an attack that focused completely on people to circumvent existing traditional security algorithms. The email appears to be a dependable, appropriate, and solid communication medium for internet users. At present, the email is submerged with spam content, both in text-based form or undesired text planted inside the images. This study reviews articles on phishing image spam classification published from 2006 to 2020 based on spam classification application domains, datasets, features sets, spam classification methods, and the measurement metrics adopted in the existing studies. More than 50 articles, both from Web of Science and Scopus databases were picked. Achieving the study's target, we carried out a broad survey and analysis to identify the domains where spam classification was applied. Furthermore, several public data sets, features set, classification methods, and measuring metrics are found and the popular once were pinpointed. The study revealed that Personal Collection, Dredze, and Spam Archives datasets are the most commonly used datasets in image spam classification research. Low-level and image metadata are the most widely used features set. The methods of image spam classification as identified in this study are supervised machine learning, unsupervised machine learning, semi-supervised machine learning, content-based and statistical learning. Among these methods, the most commonly utilized is the Support Vector Machine (SVM) which falls under supervised machine learning. This is followed by Naïve Bayes and K-Nearest Neighbor. The commonly adopted metrics for the performance evaluation of the existing image spam classifiers are also identified and briefly discussed. We compared the performance of the state-of-the-art image spam models. Lastly, we pointed out promising directions for future research.

Download

Supervised Ann Classification for Engineering Machined Textures Based on Enhanced Features Extraction and Reduction Scheme

Image classification involves the act of classifying images according to their extracted and sele... more Image classification involves the act of classifying images according to their extracted and selected features. Some of the main problems of image classification are the poor features that does not precisely represent an image, and the large dimensionality of data input passed to classifiers. To overcome these problems, an efficient feature extraction and selection technique is required which extracts and reduces the number of selected features and thus improves the classification accuracy. In this paper, feature extractions scheme followed by features dimensionality reduction technique is presented for image classification. The proposed methodology focuses mainly on three main stages for an input image, firstly extracting features by commonly used features extraction methods such as edge detection, and histogram. Secondly reducing the numbers of extracted features vector using the concept of Principal Component Analysis (PCA) for features selection and vectors dimensionality reduct...

BGLBP-based image background extraction method

Int. Arab J. Inf. Technol., 2016

Local Binary Pattern (LBP) is invariant to the monotonic changes in the grey scale domain. This p... more Local Binary Pattern (LBP) is invariant to the monotonic changes in the grey scale domain. This property enables LBP to present a texture descriptor that can be useful in applications dealing with the local illumination changes. However, the existing versions of LBP are not able to handle image illumination changes, especially in outdoor environments. These non-patterned illumination changes disturb performance of the background extraction methods. In this paper, an extended version of LBP which is called Back Ground Local Binary Pattern (BGLBP) is presented. BGLBP is designed for the background extraction application but it is extendable to the other areas as a texture descriptor. BGLBP is an extension of Direction LBP (D-LBP), Centre-Symmetric LBP (CS-LBP), Uniform Local Binary Pattern (ULBP), and RIU-LBP (RotationInvariant Uniform LBP) and it has been designed to inherit the positive properties of previous versions. The performance of BGLBP as a part of background extraction meth...

Region-based surveillance video retrieval with effective object representation

The Imaging Science Journal, 2017

ABSTRACT Retrieving the most relevant video frames that contain the object specified in a given q... more ABSTRACT Retrieving the most relevant video frames that contain the object specified in a given query (query-by-region) remains a challenging task. Two common challenges of region-based retrieval approaches are to accurately extract or segment object(s) and select a proper matching strategy. This paper addresses these problems by proposing a retrieval approach that uses a new region-based matching technique equipped with an effective object representation method. In the first stage, the proposed approach selects the most informative instances of each object that appeared in the video by utilizing an adapted clustering algorithm over the extracted features. In the retrieval stage, the new matching technique returns the most relevant sequences of video by mapping a given region with those identified representative instances of objects based on their similarity scores. The proposed approach is evaluated on standard datasets and the results demonstrate a 31% improvement in the retrieval performance compared to other state-of-the-art methods.

A Review of Ultrasound and Computed Tomography Registration Approaches

2014 International Conference on Computer Assisted System in Health, 2014

Ultrasound is widely used in minimally invasive cardiac procedures due to its convenience of use ... more Ultrasound is widely used in minimally invasive cardiac procedures due to its convenience of use and noninvasive nature. However, the low quality of ultrasound images often limits their utility as a means for guiding procedures, since it is often difficult to relate the images to their anatomical context. To improve the interpretability of the ultrasound images while maintaining ultrasound as a flexible anatomical and functional real time imaging modality, there is need for some registration techniques that integrate them with their correspond context in high quality pre-operative models such as magnetic resonance imaging or computed tomography images. It is a challenging and remarkable step as through registration, the combined information from multi-modal image acquisition systems such as ultrasound and computed tomography can be obtained by the medical practitioner for better physiological understanding, effective image guidance surgery, treatment, monitoring and diagnostic purposes. An overview of ultrasound and computed tomography registration techniques is presented in this paper.

Twelve Anchor Points Detection by Direct Point Calculation

Facial features can be categorized it into three approaches; Region Approaches, Anchor Point (lan... more Facial features can be categorized it into three approaches; Region Approaches, Anchor Point (landmark) Approaches and Contour Approaches. Generally, anchor points approach provide more accurate and consistent representation. For this reason, anchor points approach has been chose to utilize. Although, as the experiment data sets have become larger, algorithms have become more sophisticated even if the reported recognition rates are not as high as in some earlier works. This will cause a higher complexity and computer burden. Indirectly, it also will affect the time for real time face recognition systems. Here, it is proposed the approach of calculating the points directly from the text file to detect twelve anchor points ( nose tip, mouth centre, right eye centre, left eye centre, upper nose and chin). In order to get the anchor points, points for the nose tip have to be detected first. Then the upper nose and face point is localization. Lastly, the outer and inner eyes corner is localized. An experiment has been carried out with 420 models taken from GavabDB in two positions with frontal view and variation of expressions and positions. Our results are compared with three researchers that is similar to and show that better result is obtained with a median error of the eight points is around 5.53mm.

Download

An Efficient Non-Occluded Face Area Detection Method under Unconstrained Environment of Smartphone Database

2019 IEEE International Conference on Signal and Image Processing Applications (ICSIPA), 2019

Occlusion is one of the most challenging problems for recognizing faces captured in unconstrained... more Occlusion is one of the most challenging problems for recognizing faces captured in unconstrained conditions. This paper focuses on direct detect the useful part in the face which used for features extraction process. To solve occlusion problem and detect un occluded face parts, a Non-Occluded Face Area Detection (NOFAD) method based on geometric features is proposed. The performance of the proposed method is evaluated on the SFV and MOBIO databases, and the experimental results show the outperformance of the proposed method in extracting the non-occluded facial area, realising an accuracy rate of 99.4% and 99.1%, respectively.

Download

Multi-Granularity Semantic Information Integration Graph for Cross-Modal Hash Retrieval

IEEE access, 2024

With the development of intelligent collection technology and popularization of intelligent termi... more With the development of intelligent collection technology and popularization of intelligent terminals, multi-source heterogeneous data are growing rapidly. The effective utilization of rich semantic information contained in massive amounts of multi-source heterogeneous data to provide users with highquality cross-modal information retrieval services has become an urgent problem to be solved in the current field of information retrieval. In this paper, we propose a novel cross-modal retrieval method, named MGSGH, which deeply explores the internal correlation between data of different granularities by integrating coarse-grained global semantic information and fine-grained scene graph information to model global semantic concepts and local semantic relationship graphs within a modality respectively. By enforcing crossmodal consistency constraints and intra-modal similarity preservation, we effectively integrate the visual features of image data and semantic information of text data to overcome the heterogeneity between the two types of data. Furthermore, we propose a new method for learning hash codes directly, thereby reducing the impact of quantization loss. Our comprehensive experimental evaluation demonstrated the effectiveness and superiority of the proposed model in achieving accurate and efficient cross-modal retrieval.

Download

Reviewing Vegetation Indices for Mobile Application: Potentials and Challenges

Journal of Advanced Research in Applied Sciences and Engineering Technology, Dec 20, 2023

Vegetation are quantitative metrics that are used to assess the density of vegetation as well as ... more Vegetation are quantitative metrics that are used to assess the density of vegetation as well as its health and vitality. Vegetation indices are important in precision agriculture and monitoring environmental conditions. For years, vegetation indices have been derived from remote sensing or aerial imaging data, with focus on large plantation estate and green coverage. Recently, there is a growing interest in using mobile phone for measuring vegetation indices at a more affordable cost. The spectral range for vegetation indices extends from 400 nm to 900 nm, which places it in the VIS and NIR spectral categories, respectively. The CMOS sensor that is typically found in mobile phone cameras has the capacity to capture NIR images. This paper aims to give a background in understanding vegetation indices available and weighing the suitability of using these indices for mobile applications. The potential of using these indices has been establish when the studies indicate several works have successfully used mobile phone camera to capture vegetation indices formulation, but still additional features is needed to ensure the standardisation and consistency of image calibration.

Download

Grain recognition based on colour and shape analysis

Download

Uploads

Papers by fatimah khalid

Log In