Pallab Kanti Roy

Retinal Image Quality Classification Using Saliency Maps and CNNs

Retinal image quality assessment (IQA) algorithms use different hand crafted features without considering the important role of the human visual system (HVS). We solve the IQA problem using the principles behind the working of the HVS. Unsupervised information from local saliency maps and supervised information from trained convolutional neural networks (CNNs) are combined to make a final decision on image quality. A novel algorithm is proposed that calculates saliency values for every image pixel at multiple scales to capture global and local image information. This extracts generalized image information in an unsupervised manner while CNNs provide a principled approach to feature learning without the need to define hand-crafted features. The individual classification decisions are fused by weighting them according to their confidence scores. Experimental results on real datasets demonstrate the superior performance of our proposed algorithm over competing methods.

Segmentation of Right Ventricle in Cardiac MR Images Using Shape Regression

Accurate and automatic segmentation of the right ventricle is challenging due to its complex anatomy and large shape variation observed between patients. In this paper the ability of shape regression is explored to segment right ventricle in presence of large shape variation among the patients. We propose a robust and efficient cascaded shape regression method which iteratively learns the final shape from a given initial shape. We use gradient boosted regression trees to learn each regressor in the cascade to take the advantage of supervised feature selection mechanism. A novel data augmentation method is proposed to generate synthetic training samples to improve regressors performance. In addition to that, a robust fusion method is proposed to reduce the the variance in the predictions given by different initial shapes, which is a major drawback of cascade regression based methods. The proposed method is evaluated on an image set of 45 patients and shows high segmentation accuracy with dice metric of

Automatic Eye Type Detection in Retinal Fundus Image Using Fusion of Transfer Learning and Anatomical Features

0.87\pm 0.06

A novel hybrid approach for severity assessment of Diabetic Retinopathy in colour fundus images

. Comparative study shows that our proposed method performs better than state-of-the-art multi-atlas label fusion based segmentation methods.

A CNN based neurobiology inspired approach for retinal image quality assessment

The fovea is one of the most important anatomical landmarks in the eye and its localization is required in automated analysis of retinal diseases due to its role in sharp central vision. In this paper, we propose a two-stage deep learning framework for accurate segmentation of the fovea in retinal colour fundus images. In the first stage, coarse segmentation is performed to localize the fovea in the fundus image. The location information from the first stage is then used to perform fine-grained segmentation of the fovea region in the second stage. The proposed method performs end-to-end pixelwise segmentation by creating a deep learning model based on fully convolutional neural networks, which does not require the prior knowledge of the location of other retinal structures such as optic disc (OD) and vasculature geometry. We demonstrate the effectiveness of our method on a dataset with 400 retinal images with average localization error of 14 ± 7 pixels.

Multi-atlas label fusion using hybrid of discriminative and generative classifiers for segmentation of cardiac MR images.

Retinal fundus images are mainly used by ophthalmologists to diagnose and monitor the development of retinal and systemic diseases. A number of computer-aided diagnosis (CAD) systems have been developed aimed at automation of mass screening and diagnosis of retinal diseases. Eye type (left or right eye) of a given retinal image is an important meta data information for a CAD. At present, eye type is graded manually which is time consuming and error prone. This article presents an automatic method for eye type detection, which can be integrated into existing retinal CAD systems to make them more faster and accurate. Our method combines transfer learning and anatomical prior knowledge based features to maximize the classification accuracy. We evaluate the proposed method on a retinal image set containing 5000 images. Our method shows a classification accuracy of 94% (area under the receiver operating characteristics curve (AUC) = 0.990).

Segmentation of optic disc and optic cup in retinal fundus images using shape regression

This paper presents an efficient and robust approach to detect right ventricular landmark points in short axis cardiac MRI, based on multiscale HOG descriptor and random forest classifier. First, candidate landmark locations are determined using multiscale Harris corner detector. Multiscale HOG descriptor is then extracted at the candidate search locations. A probabilistic random forest classifier model is trained to discriminate landmark points from non-landmark regions. The landmark position is then estimated as the weighted average of the candidate locations where weights are computed from the probability scores derived from the classifier. Experimental result performed on an image set of 15 patients demonstrates the effectiveness of our proposed method with average error (Euclidean distance between the detected landmark and the manually annotated landmark points) of 5.06 pixels. Contrary to most existing approaches, our proposed method has minor dependency to prior segmentation of right ventricle, hence is less affected by plausible segmentation error.

Image Quality Classification for DR Screening Using Convolutional Neural Networks

Diabetic Retinopathy (DR) is one of the leading causes of blindness worldwide. Detecting DR and grading its severity is essential for disease treatment. Convolutional neural networks (CNNs) have achieved state-of-the-art performance in many different visual classification tasks. In this paper, we propose to combine CNNs with dictionary based approaches, which incorporates pathology specific image representation into the learning framework, for improved DR severity classification. Specifically, we construct discriminative and generative pathology histograms and combine them with feature representations extracted from fully connected CNN layers. Our experimental results indicate that the proposed method shows improvement in quadratic kappa score (κ2 = 0.86) compared to the state-of-the-art CNN based method (κ2 = 0.81).