Semantic Segmentation of Periocular Near-Infra-Red Eye Images Under Alcohol Effects

Abstract

This paper proposes a new framework to detect, segment, and estimate the localization of the eyes from a periocular Near-Infra-Red iris image under alcohol consumption. This stage will take part in the final solution to measure the fitness for duty. Fitness systems allow us to determine whether a person is physically or psychologically able to perform their tasks. Our segmentation framework is based on an object detector trained from scratch to detect both eyes from a single image. Then, two efficient networks were used for semantic segmentation; a Criss-Cross attention network and DenseNet10, with only 122,514 and 210,732 parameters, respectively. These networks can find the pupil, iris, and sclera. In the end, the binary output eye mask is used for pupil and iris diameter estimation with high precision. Five state-of-theart algorithms were used for this purpose. A mixed proposal reached the best results. A second contribution is establishing an alcohol behavior curve to detect the alcohol presence utilizing a stream of images captured from an iris instance. Also, a manually labeled database with more than 20k images was created. Our best method obtains a mean Intersection-over-Union of 94.54% with DenseNet10 with only 210,732 parameters and an error of only 1-pixel on average.INDEX TERMS Biometrics, Fitness for duty, segmentation, iris, alcohol.

Exploring Bias in Sclera Segmentation Models: A Group Evaluation Approach

“…While these techniques provided the basis for early sclerabased biometric systems, recent work is looking increasingly at deep learning models that have been shown to provide excellent segmentation performance for highly diverse input images. Examples of techniques from this group include convolutional neural networks (CNNs) with an encoder-decoder design [5], [9], [44], fully convolutional models [45], densely connected convolutional networks [46], generative networks [1], and other (custom) deep learning approaches [47], [48].…”

Toward an Efficient Iris Recognition System on Embedded Devices

“…Recent works in IR use deep learning to segment and localise the pupil and the iris in a periocular image [11], [31]- [33]. The Criss-Cross Attention Network (CCNet), developed by Mishra et al [33], is an iris semantic segmentation network based on U-Net [32].…”

“…Some of the proposed approaches used classical methods of image processing, such as Gabor Wavelets, [1]- [3], Localized Binary Patterns (LBP) [4], or Binarized Statistical Image Features (BSIF) [5]. In contrast, others used Convolutional Neural Networks (CNN) for segmentation and/or encoding [7]- [11]. However, only a few previous works aimed to describe how to build a binocular iris imaging device efficiently from scratch at both hardware and software levels.…”

“…• Efficiency: The proposed architecture of UNet_xxs reduces the parameter number by 77% with respect to State of the art (CCNet). This makes UNet_xxs faster and easier to implement than previous works [11]. Additionally, joining available hardware and software components into a single cohesive system with a practical use case is another innovative point.…”

Behavioural Curves Analysis Using Near-Infrared-Iris Image Sequences

“…The eye detector module was implemented to find both eyes in the input periocular images and for subsequent cropping and segmentation. This algorithm is detailed in our previous work on semantic segmentation of periocular NIR images under alcohol effects [54]. This module applied Eye-tiny-yolo, classical tracking, and semantic segmentation by Cluster-Coordinated Net [55] (CCNet).…”