Browsing by Author "Smitha, A."

Now showing 1 - 10 of 10

A nonlocal deep image prior model to restore optical coherence tomographic images from gamma distributed speckle noise
(Taylor and Francis Ltd., 2021) Smitha, A.; Padikkal, P.
Optical Coherence Tomography (OCT) is often employed to observe the retinal layers in the human eyes. The retinal scans are susceptible to artefacts such as head movements or eye blinks. Along with this, the quality of the images is degraded by speckle noise caused due to the constructive and destructive interference of the waves used for capturing data. Recently, image restoration techniques have geared up in terms of quality with the exertion of deep learning. Despeckling using deep learning, in general, necessitates a large set of training images. On the contrary, deep image prior is a novel model that performs denoising operations using a single training image, based on a prior assumption about the noise distribution. This paper extends the concept of the deep image prior towards non-local restoration for speckle noise assuming that the speckle follows Gamma distribution. Such a framework can be incorporated to enhance the OCT images. The proposed framework is assessed qualitatively with visual comparisons and quantitatively using statistical measures like PSNR, CNR and ENL. Comparative studies confirm that the proposed method outperforms the existing methods in restoring speckled input images. © 2021 Informa UK Limited, trading as Taylor & Francis Group.
A retinex based non-local total generalized variation framework for OCT image restoration
(Elsevier Ltd, 2022) Smitha, A.; Febin, I.P.; Padikkal, P.
A retinex driven non-local total generalized variational (TGV) model is proposed in this paper to restore and enhance speckled images. The combined first and second-order TGV controlled by a balancing parameter are used to improve the enhancement and restoration process. The distribution of the speckle is estimated from input images using detailed statistical analysis. The model is designed to handle speckle-noise following a Gamma distribution, as analyzed later in this paper. The non-local TGV model is shown to restore images without causing any visual artefacts, unlike the normal total variation (TV) model. Moreover, a retinex framework shows a remarkable improvement to the contrast features of the data without distorting the natural image characteristics as quantified visually and statistically in the experimental section of this work. A fast numerical approximation based on the Split-Bregman scheme is employed to improve the efficiency of the model in terms of computation. The proposed model is verified to have despeckled and enhanced the Optical Coherence Tomography (OCT) data to a greater extent compared to the state-of-the-art models as observable from the results shown in this paper. © 2021 Elsevier Ltd
A Semi-supervised Generative Adversarial Network for Retinal Analysis from Fundus Images
(Springer Science and Business Media Deutschland GmbH, 2021) Smitha, A.; Jidesh, P.
Retinal disorders are the prominent diseases causing visual impairments to a large population across the globe. A prompt diagnosis can address this problem to a large extend. Moreover, AI-enabled devices help ophthalmologists in the timely diagnosis of the disorders and initiate appropriate treatment. In this work, we develop a GAN based semi-supervised model to extract the prominent retinal structures and classify the fundus images as normal or abnormal using data from multiple repositories. The dice coefficient of 0.9 across various datasets affirms the good performance on source independent data. Such a model can be extended to incorporate additional features and be integrated into the ophthalmoscopes for quick retinal examination through telemedicine. Â© 2021, Springer Nature Singapore Pte Ltd.
Assessing Damage ofÂ Natural Disasters fromÂ Satellite Imagery Using aÂ Deep Learning Model
(Springer Science and Business Media Deutschland GmbH, 2023) Tikle, S.; Jidesh, P.; Smitha, A.
Natural disasters are events that arise anywhere on the planet. It causes enormous devastation and places entire cities in need of significant support. The capability to swiftly and precisely deploy rescue services in the affected regions is critical for reducing the impact and saving lives. A two-step model is developed in an attempt to resolve this problem by using satellite images as input. The model draws attention to the structures such as buildings, which are severely damaged. The current deep learning-based computer vision models use pre- and post-disaster satellite images to semantically infer the level of damage to individual buildings after natural disasters. This model alleviates an important roadblock in disaster managerial decisions by simplifying the evaluations of damages caused to the building. We used DeepLabv3+ for semantic segmentation and a custom CNN model for image classification to analyze disaster-related images. This paper describes how satellite data and proficient image analysis may effectively be used to conduct disaster and crisis management to assist jobs that require fast mappings. This modelâ€™s performance and accuracy are sub-optimal and are being studied to further improvisations. However, it surpasses the current cutting-edge model. Â© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Classification of Multiple Retinal Disorders from Enhanced Fundus Images Using Semi-supervised GAN
(Springer, 2022) Smitha, A.; Padikkal, P.
Automatic detection of retinal disorders is gaining considerable attention with the emergence of deep learning. Ophthalmologists primarily use color fundus photographs to examine the human retina and diagnose the abnormalities. As there is a surge in the number of visual impairments, an AI-enabled retina screening system can expedite the retina examination process. Existing works in this direction are primarily focused on either segmentation or classification. Furthermore, the majority of the works are implemented using preprocessed good quality fundus images. In reality, however, the quality of color fundus images is degraded due to the illumination inhomogeneity and low contrast issues. Thus, there is a need to develop an end-to-end fundus image analysing system. Steering in this direction, the proposed work attempts to analyze the performance of semi-supervised Generative Adversarial Networks (GANs) for the classification of retinal fundus images into multiple categories. Besides, the nonlocal retinex framework is applied to enhance the quality of fundus images without over-smoothing the edges. The large data set of raw fundus acquired from multiple Eye hospitals and released in public domain is used to implement the proposed work. The results obtained are compared with the transfer learning method, and an average accuracy of 87% is obtained. It suggests that the semi-supervised GANs can be potentially used to classify heterogeneous retinal disorders. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2021.
Comparative Analysis ofÂ Machine Learning andÂ Deep Learning Models forÂ Ship Classification fromÂ Satellite Images
(Springer Science and Business Media Deutschland GmbH, 2022) Hazarika, H.; Jidesh, P.; Smitha, A.
The automatic detection of the ship from satellite image analysis is the limelight of research in recent years due to its widespread applications. In this paper, a handful of traditional machine learning and deep learning models are compared based on their performance to classify the satellite images available in the public repository as a ship or other categories. The Support Vector Machine(SVM), Decision Trees, Random Forest, K-Nearest Neighbor (KNN), Gaussian Naive Bayes (GaussianNB), and Logistic Regression are machine learning models used in the present work. Histogram of Gradient (HoG) features are used as feature descriptors considering the diverse size and shape of ships in the satellite image dataset. Transfer learning is applied using the deep learning models namely, Inception and ResNet, that are fine-tuned for various learning rates and optimizers. The meticulous experimentation carried out reveals that traditional machine learning performs well when trained and tested on a single dataset. However, there is a drastic change in the performance of machine learning models when tested on a different ship dataset. The results show that the deep learning models have better feature detection and thus have better performance when transfer learning is used on various datasets. Â© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Detection of retinal disorders from OCT images using generative adversarial networks
(Springer, 2022) Smitha, A.; Padikkal, J.
Retinal image analysis has opened up a new window for prompt diagnosis and detection of various retinal disorders. Optical Coherence Tomography (OCT) is one of the major diagnostic tools to identify retinal abnormalities related to macular disorders like Age-Related Macular Degeneration (AMD) and Diabetic Macular Edema (DME). The clinical findings include retinal layer analysis to spot the abnormalities on OCT images. Though various models are proposed over the years to diagnose these disorders automatically, an end-to-end system that performs automatic denoising, segmentation, and classification does not exist to the best of our knowledge. This paper proposes a Generative Adversarial Network (GAN) based approach for automated segmentation and classification of OCT-B scans to diagnose AMD and DME. The proposed method incorporates the integration of handcrafted Gabor features to enhance the retina layer segmentation and non-local denoising to remove speckle noise. The classification metrics of GAN are compared with existing methods. The accuracy of up to 92.42% and F1-score of 0.79 indicates that the GANs can perform well for segmentation and classification of OCT images. © 2022, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
Development and validation of a novel automated method for quantification of choroidal thickness in age-related macular degeneration
(SPIE, 2021) Smitha, A.; Jidesh, P.; Janarthanam, J.; Lakshminarayanan, V.
Age-related Macular Degeneration (AMD) is a progressive, irreversible retinal disorder, and one of the leading causes of severe visual impairment or even blindness in the elderly population. The choroid plays a vital role in the pathophysiology of AMD. It is known that abnormal choroidal blood flow leads to retinal photoreceptor dysfunction and eventual death. We propose a new automated algorithm that can be used to quantify choroidal thickness (CT) from Optical Coherence Tomography (OCT) images of the retina. This thickness evaluation procedure includes image contrast enhancement, localization around the fovea centralis, segmentation of Retinal Pigment Epithelium (RPE) and choroidal layer, followed by CT measurement at multiple locations in the sub-foveal region at intervals of 0.5 mm on both nasal and temporal sides up to a distance of 1 mm from the center of the foveal pit. The horizontal radial scan OCT images (Cirrus 5000, Carl Zeiss Meditec Inc., Dublin, CA) of both healthy and AMD patients were used to measure the CT using the new algorithm. The statistical tests convey that the CT of AMD patients is relatively smaller than the normal condition. Furthermore, t-Test conducted between the proposed approach and clinical approach of extracting CT measurements confirm that the proposed method is in good agreement with the clinical measurements. On an average, the thickness of the choroid is found to be 0.32 Â± 0.10 mm for the normal category and 0.21 Â± 0.06 mm for the AMD category, in the central sub-foveal region, as obtained from the proposed automatic CT measurement method. The clinical significance and the results of automated choroid extraction are discussed in this paper. Â© COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.
Restoration and Enhancement of Aerial and Synthetic Aperture Radar Images Using Generative Deep Image Prior Architecture
(Springer Science and Business Media Deutschland GmbH, 2022) Shastry, A.; Smitha, A.; George, S.; Padikkal, J.
Restoration and enhancement of low light images is an inevitable pre-processing activity among remote sensing, aerial and satellite imaging modalities. The images captured under various atmospheric conditions are distorted. Therefore, they need a thorough conditioning before being analysed. In this paper, we propose a retinex-based variational framework designed under a generative deep image prior architecture to restore and enhance distorted images from satellite, aerial and remote sensing applications. The model handles data-correlated speckle noise found in active image sensing modalities, duly considering its distribution. The data-fidelity aspect of the proposed variational framework is designed using the Bayesian Maximum A Posteriori (MAP) estimate, assuming that the input images are contaminated with Gamma distributed speckled interference. Further, model is catered to handle various noise distributions, such as Gaussian and Poisson, by appropriately altering the data fidelity term specific to the distribution, without modifying the architecture of the model. The variational retinex model employed herein also addresses contrast degradation and intensity inhomogeneity aberrations in the input images. The proposed model is assessed qualitatively using visual comparisons and quantified using the relevant statistical measures. The experimental results confirm that the proposed model outperforms the existing methods in terms of restoration and contrast enhancement of speckled images. The proposed method also has shown the full potential to adapt the model to restore the degraded images following any distribution. © 2022, Deutsche Gesellschaft für Photogrammetrie, Fernerkundung und Geoinformation (DGPF) e.V.
Retinal Vessel Classification Using the Non-local Retinex Method
(Springer, 2020) Smitha, A.; Jidesh, P.; Febin, I.P.
Automatic retinal vessel segmentation has turned out to be highly propitious for medical practitioners to diagnose diseases like glaucoma and diabetic retinopathy. These diseases are classified based on the thickness of the retinal vessel, the pressure imposed on the nerve endings and optical disc to cup ratio of the retina. The state-of-the-art device for this purpose presently available in the market is expensive and has scope to meliorate sensitivity and precision of its performance. Thus, automatic retinal blood vessel segmentation and classification is the need of the hour. In this paper, a novel non-local total variational retinex based retinal image preprocessing approach is proposed to extract the retinal vessel features and classify the vessels using ground truth images. Matlab implementation results indicate that an average accuracy of 94% with an acceptable range of sensitivity and specificity could be achieved on the retinal image database available online. Â© 2020, Springer Nature Switzerland AG.