IRIS TUMOR DETECTION USING CNN
Abstract : Iris tumors, so called intraocular tumors are kind of tumors that start in the iris; the colored part of the eye that surrounds the pupil. There is a need for an accurate and cost-effective iris tumor detection system since the available techniques used currently are still not efficient. The combination of the image processing different techniques has a great efficiency for the diagnosis and detection of the iris tumor. Image processing techniques improve the diagnosis of the tumor by enhancing the quality of the images, so the physicians diagnose properly. Moreover, using some techniques such as edge detection and image fusion helps in detecting and segmenting the tumor, located in the iris. This paper aims to develop a detection system that automatically detects the presence of abnormalities or tumors in the iris. The suggested system combines different image processing techniques such as image filtering, images adding, canny edge detection, and image fusion. These methods are used in order to analyze and segment the tumor into the iris, and then mark this abnormal region onto the original grayscale image. The iris images are obtained from a public database available on the internet, “Miles Research”. The computer experimental results proved that the proposed detection system can successfully segment an iris tumor, and mark it to be then superimposed on the original image using image fusion.
• The survey of various existing techniques provides a platform for the development of the novel techniques in this area as future work.
• Existing iris recognition systems are heavily dependent on specific conditions, such as the distance of image acquisition and the stop-and-stare environment, which require significant user cooperation.
• Most existing iris recognition algorithms are designed for highly controlled cooperative environments, which is the cause of their failure in non-cooperative environments, i.e., those that include noise, off-angles, motion blurs, glasses, hairs, specular reflection (SR), eyelids and eyelashes incorporation, and partially open eyes.
• If the dark pixel exists inside the box because the dark pupil can be included in the box, the box position is adaptively moved into the lower direction until it does not include the dark pixel, and the average RGB value is extracted box.
• When applying CNN, false positive errors (non-iris pixel is incorrectly classified into iris pixel) exist in the pupil area.
• Iris DenseNet is an end-to-end segmentation network that uses the complete image without prior pre-processing or other conventional image processing techniques with the best information gradient flow, which prevents the network from overfitting and vanishing gradient problem.
• To address the issues of accurate segmentation without prior pre-processing and to develop a robust scheme for all types of environments, this study presents a densely connected fully convolutional network (IrisDenseNet)-based approach to detect an accurate iris boundary with better information gradient flow due to dense connectivity.
• To address the iris segmentation issues in challenging situations by visible light and near-infrared light camera sensors, this paper proposes a densely connected fully convolutional network (IrisDenseNet), which can determine the true iris boundary even with inferior-quality images by using better information gradient flow between the dense blocks.
• To address this issue, we proposed a two-stage iris segmentation method based on convolutional neural networks (CNN), which is capable of robustly finding the true iris boundary in the above-mentioned intense cases with limited user cooperation.
• Our proposed iris segmentation scheme can be used with inferior quality noisy images even in visible light environment.
• The proposed method accurately identified the true boundary even in intense scenarios, such as glasses, off-angle eyes, rotated eyes, side view, and partially opened eyes.
• We proposed a CNN-based segmentation method of the iris region based on the ROI defined by the rough iris boundary.
• Our proposed method can correctly segment the iris region irrespective of various noises in eye image.
• The major challenge here is to achieve high performance on a mobile platform because of limitation of space, power, and cost of the system.
• IrisDenseNet is a densely connected fully convolutional network that proceeds in the feed-forward fashion with dense features for better performance.
• This connectivity enhances the capability of the network and enables the feature reuse for better performance.
• Biometrics in both physiological and behavioral forms are delivering an efficient platform for security metrics.
• We would optimize the network further and reduce the number of layers to make it memory-efficient for mobile and handheld devices with reduced parameters and multiplications.
• Iris recognition shares the following advantages of secure biometrics: iris features are unique even in the case of twins, left eye iris features of an individual are different from right eye iris features, iris features are naturally complex to be created artificially, and iris features are permanent and remain same throughout a human’s lifespan
|