Dr. Deepthy Mary Alex

RESEARCH, TEACHING, or OTHER INTERESTS

Engineering, Computer Vision and Pattern Recognition, Artificial Intelligence, Signal Processing

Scopus Publications

DANNET: deep attention neural network for efficient ear identification in biometrics
Deepthy Mary Alex, Kalpana Chowdary M., Hanan Abdullah Mengash, Venkata Dasu M., Natalia Kryvinska, Chinna Babu J., and Ajmeera Kiran
PeerJ
Biometric identification, particularly ear biometrics, has gained prominence amidst the global prevalence of mask-wearing, exacerbated by the COVID-19 outbreak. This shift has highlighted the need for reliable biometric systems that can function effectively even when facial features are partially obscured. Despite numerous proposed convolutional neural network (CNN) based deep learning techniques for ear detection, achieving the expected efficiency and accuracy remains a challenge. In this manuscript, we propose a sophisticated method for ear biometric identification, named the encoder-decoder deep learning ensemble technique incorporating attention blocks. This innovative approach leverages the strengths of encoder-decoder architectures and attention mechanisms to enhance the precision and reliability of ear detection and segmentation. Specifically, our method employs an ensemble of two YSegNets, which significantly improves the performance over a single YSegNet. The use of an ensemble method is crucial in ear biometrics due to the variability and complexity of ear shapes and the potential for partial occlusions. By combining the outputs of two YSegNets, our approach can capture a wider range of features and reduce the risk of false positives and negatives, leading to more robust and accurate segmentation results. Experimental validation of the proposed method was conducted using a combination of data from the EarVN1.0, AMI, and Human Face datasets. The results demonstrate the effectiveness of our approach, achieving a segmentation framework accuracy of 98.93%. This high level of accuracy underscores the potential of our method for practical applications in biometric identification. The proposed innovative method demonstrates significant potential for individual recognition, particularly in scenarios involving large gatherings. When complemented by an effective surveillance system, our method can contribute to improved security and identification processes in public spaces. This research not only advances the field of ear biometrics but also provides a viable solution for biometric identification in the context of mask-wearing and other facial obstructions.


Retinal Image Enhancement based on illumination component and gamma correction
S. P. James, D. A. Chandy, and D. M. Alex
Institution of Engineering and Technology



YSegNet: a novel deep learning network for kidney segmentation in 2D ultrasound images
Deepthy Mary Alex, D. Abraham Chandy, A. Hepzibah Christinal, Arvinder Singh, and M. Pushkaran
Springer Science and Business Media LLC



A Hybrid Random Forest Classifier for Chronic Kidney Disease Prediction from 2D Ultrasound Kidney Images
Deepthy Mary Alex, D. Abraham Chandy, A. Hepzibah Christinal, Arvinder Singh, and M. Pushkaran
World Scientific Pub Co Pte Ltd
Chronic kidney disease (CKD) is one of the causes of mortality in almost all countries across the globe and the notable thing is its asymptomatic nature in the early stages. This disease is characterized by the gradual loss of kidney function in an individual. Frequently chronic kidney disease is diagnosed based on the Estimated Glomerular Filtration Rate (eGFR) determined from blood and urine tests. In order to reduce the risk factors arising due to chronic kidney disease, it is essential to be diagnosed in the earlier stages itself. This work proposes an automated chronic kidney disease detection based on the textural features of the kidney using a hybrid random forest classifier from 2D ultrasound kidney images. The proposed classifier is compared with the other competing machine learning classifiers through experimenting on a dataset of 150 images and gives a better accuracy of [Formula: see text] with [Formula: see text] of recall and precision, thus proving it to be promising in detecting CKD noninvasively in the early stages.


Exploration of a framework for the identification of chronic kidney disease based on 2d ultrasound images: A survey
Deepthy Mary Alex and D. Abraham Chandy
Bentham Science Publishers Ltd.
Background: Chronic kidney disease (CKD) is a fatal disease that ultimately results in kidney failure. The primary threat is the aetiology of CKD. Over the years, researchers have proposed various techniques and methods to detect and diagnose the disease. The conventional method of detecting CKD is the determination of the estimated glomerular filtration rate by measuring creatinine levels in blood or urine. Conventional methods for the detection and classification of CKD are tedious; therefore, several researchers have suggested various alternative methods. Recently, the research community has shown keen interest in developing methods for the early detection of this disease using imaging modalities such as ultrasound, magnetic resonance imaging, and computed tomography. Discussion: The study aimed to conduct a systematic review of various existing techniques for the detection and classification of different stages of CKD using 2D ultrasound imaging of the kidney. The review was confined to 2D ultrasound images alone, considering the feasibility of implementation even in underdeveloped countries because 2D ultrasound scans are more cost effective than other modalities. The techniques and experimentation in each work were thoroughly studied and discussed in this review. Conclusion: This review displayed the cutting-age research, challenges, and possibilities of further research and development in the detection and classification of CKD.


Evaluation of Inpainting in Speckled and Despeckled 2D Ultrasound Medical Images
Deepthy Mary Alex and D. Abraham Chandy
IEEE
Image processing has paved its way through various applications in the medical field, where researches are being carried out to prompt and improvise the care and treatment given to patients in all aspects. Medical image processing applications concerns the need for obtaining images as per the requirement. Inpainting of markers on medical images generated using various modalities is crucial to process the images for numerous applications. One of the inherent and profoundly seen problem in medical images, especially in the ultrasound images is the speckle noise. It alters the edges, fine details and other features of organs that is required for interpretation. In this paper, the need for despeckling of 2D ultrasound images is studied by evaluating the performance metrics for inpainting based on fast marching algorithm of speckled and despeckled 2D ultrasound images.


Speckle noise suppression in 2D ultrasound kidney images using local pattern based topological derivative
Deepthy Mary Alex, A. Hepzibah Christinal, D. Abraham Chandy, Arvinder Singh, and M Pushkaran
Elsevier BV



Morphology based inpainting for removal of markers in medical images




Investigations on performances of pre-trained u-net models for 2d ultrasound kidney image segmentation
Deepthy Mary Alex and D. Abraham Chandy
Springer International Publishing



Comparison of PSO and Sequential Search Algorithms for Improvisation of Entropy-Based Ear Localization
D. Abraham Chandy, Hepzibah A. Christinal, Rakhi Chandrasekaran, and Deepthy Mary Alex
Springer Singapore



Rationalizing of Morphological Renal Parameters and eGFR for Chronic Kidney Disease Detection
Deepthy Mary Alex, D. Abraham Chandy, and Anand Paul
Chapman and Hall/CRC