Periocular Region

About: Periocular Region is a research topic. Over the lifetime, 256 publications have been published within this topic receiving 4424 citations.

Convolutional Neural Networks and Periocular Region Image Recognition

Abstract: There are some benefits in using periocular biometric traits for individual identification. This work describes the use of convolutional neural network Neocognitron, in this novel application, in individual recognition using periocular region images. Besides, it is used the competitive learning using the extreme points of lines detected in the preprocessing of the input images as winner positions. It was used Carnegie Mellon University - Pose, Illumination, and Expression Database (CMU-PIE), with 41,368 images of 68 persons. From these images, 57 \(\times \) 57 periocular images were obtained as training and test samples. The experiments indicate results in the Kappa index of 0.89, for periocular images, and 0.91 for complete face images.

Surgical treatment of eyelid tumors

Abstract: Eyelid tumors are often periocular skin lesions but can also be derived from adnexal structures or the conjunctiva. Treatment is primarily surgical, especially for malignant tumors, whereby complete excision is strived for to minimize the local rate of recurrence. Adjuvant therapy and conservative alternative methods (e.g. radiotherapy, cryotherapy and immunomodulation) are reserved for specific cases. The main goal for reconstruction of the eyelids is the functional and aesthetic rehabilitation. This article focuses on the principles and underlying techniques for the reconstruction of the periocular region.

Weighted comparison score fusion for accurate verification of surgically altered periocular region

Abstract: Facial surgeries are inevitably increasing in the modern society due to factors such as consciousness for beauty/appearance either to correct the deformed structure as a result of the injury, trauma or natural process of aging. These surgical changes in the facial structure result in modified appearance of a person, which cannot be modelled using morphological or mathematical approaches. These changes directly impact the biometric performance of a recognition system, when a subject is enrolled prior to surgical modifications and is verified after the surgery. Emphasizing on the periocular region as a stand-alone biometric characteristic is on one hand increasing while on the other hand, in the target population the number of surgeries for periocular region is also increasing. These surgical changes have a direct negative impact on the biometric performance. In this work, we evaluate the degraded performance as a result of such surgeries in the periocular region by enrolling the 402 unique periocular images prior to surgery and verify mated images post-surgery by employing various well-known feature extraction techniques. The obtained verification performance equals GMR of 85.32% at a FMR of 0.01% indicating the challenge for a real-life application. Further, we explore weighted comparison score level fusion in which the weights are determined using Fischer Discriminant Ratio. The proposed weighted score level fusion in this work provides an improved verification performance of GMR = 100% at FMR = 0.01% signifying the robust nature of fused scores as compared to state-of-art verification schemes in handling the challenging data arising due to surgical modifications.

Simultaneous Iris and Periocular Region Detection Using Coarse Annotations

Abstract: In this work, we propose to detect the iris and periocular regions simultaneously using coarse annotations and two well-known object detectors: YOLOv2 and Faster R-CNN. We believe coarse annotations can be used in recognition systems based on the iris and periocular regions, given the much smaller engineering effort required to manually annotate the training images. We manually made coarse annotations of the iris and periocular regions (122K images from the visible (VIS) spectrum and 38K images from the near-infrared (NIR) spectrum). The iris annotations in the NIR databases were generated semi-automatically by first applying an iris segmentation CNN and then performing a manual inspection. These annotations were made for 11 well-known public databases (3 NIR and 8 VIS) designed for the iris-based recognition problem and are publicly available to the research community. Experimenting our proposal on these databases, we highlight two results. First, the Faster R-CNN + Feature Pyramid Network (FPN) model reported an Intersection over Union (IoU) higher than YOLOv2 (91.86% vs 85.30%). Second, the detection of the iris and periocular regions being performed simultaneously is as accurate as performed separately, but with a lower computational cost, i.e., two tasks were carried out at the cost of one.

composition for eyelid or periocular region washing and pad comprising thesame

Abstract: The present invention relates to a composition for cleaning eyelids or periocular regions, and a pad comprising the same. The present invention can be used to remove deposited wastes from eyelids and periocular regions, protect the skin from external stimulation elements, and prevent blepharitis, Demodex, etc., and the pad comprising the composition for cleaning eyelids or periocular regions is easy to use because only necessary parts can be wiped without the need for separate cleaning.