Cross-eyed 2017: Cross-spectral iris/periocular recognition competition

Sequeira, Ana F.; Chen, Lulu; Ferryman, James; Wild, Peter; Alonso‐Fernandez, Fernando; Bigün, Josef; Raja, Kiran B.; Raghavendra, R.; Busch, Christoph; Pereira, Tiago de Freitas; Marcel, Sébastien; Behera, Sushree S.; Gour, Mahesh; Kanhangad, Vivek

doi:10.1109/btas.2017.8272762

Cited by 35 publications

(27 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such DCNN, with no adaptation, presents an average rank one recognition rate of 73.80%. Adapting only [29] 96.9% (1.3) Repro-GFK in [30] 93.3% (1.4) ducible MLBPs + DoG features in [18] 62.3% ( We also observed the same trends using Triplet Networks as a base trainer. Adapting θ t[1−1] the average rank one recognition rates is improved ≈ 74%.…”

Section: Cuhk Cufs (Vis-sketch)supporting

confidence: 60%

“…An Equal Error Rate of 1.65% was achieved under the Cross-Spectral Iris/Periocular Recognition Competition [30].…”

Section: Feature Learning Based Methodsmentioning

confidence: 96%

“…Based on Geodesic Flow Kernels (GFK) the work in [30] models D s and D t in separated d-dimensional linear subspaces (φ s and φ t ) and embeds them onto a Grassmann manifold. Then, an explicit map between these subspaces is built (called Geodesic Flow); such map is encoded in a kernel G. The comparison between projected samples from the source domain (x s ) and target domain (x t ) is carried out by the kernalized dot product φ…”

Section: Feature Learning Based Methodsmentioning

confidence: 99%

“…(Table 2) 23.70% (1.89) -CDFL in [28] (Table I) 71.5% (1.4) 55.10% Paper DSIFT in [46] (Table II) 73.28% (1.10) -lines FaceVACS in [46] (Table I) 58.56% (1.19) -Gabor+RBM [28] (Table I) 86.1% (0.1) 81.29%(1.82) TRIVET in [45] 95.74% (0.5) 91.03% (1.3) IDR in [57] 95.82% (0.76) 94.03% (1.06) CDL in [58] 98.62% (0.2) 98.32% (0.05) WCNN in [59] 98.7% (0.3) 98.4% (0.4) LBPs+DoG feat. in [18] 70.33% (1.2) 74.1% (0.5) Repro-ISV in [29] 72.39% (1.35) 72.4% (0.5) ducible GFK in [30] 26.98% (0.9) 32.9% (1.) baselines Siam.…”

Section: Casia (Vis-nir)mentioning

confidence: 99%

“…We can suggest that there are domain specific feature detectors and such feature detectors need to be taken in to account for the HF R task. [18] 19.8% (1.3) Repro-ISV in [29] 89.3% (0.8) ducible GFK in [30] 17.2% (1.7) baselines Siam. Inception Resnet v2 θ t [1−5] 92.8%(1.…”

Section: Nivl (Vis-nir)mentioning

confidence: 99%

See 4 more Smart Citations

Heterogeneous Face Recognition Using Domain Specific Units

Pereira

Anjos

Marcel

2019

IEEE Trans.Inform.Forensic Secur.

Self Cite

View full text Add to dashboard Cite

The task of Heterogeneous Face Recognition consists in matching face images that are sensed in different domains, such as sketches to photographs (visual spectra images), thermal images to photographs or near-infrared images to photographs. In this work we suggest that high level features of Deep Convolutional Neural Networks trained on visual spectra images are potentially domain independent and can be used to encode faces sensed in different image domains. A generic framework for Heterogeneous Face Recognition is proposed by adapting Deep Convolutional Neural Networks low level features in, so called, "Domain Specific Units". The adaptation using Domain Specific Units allow the learning of shallow feature detectors specific for each new image domain. Furthermore, it handles its transformation to a generic face space shared between all image domains. Experiments carried out with four different face databases covering three different image domains show substantial improvements, in terms of recognition rate, surpassing the state-of-the-art for most of them. This work is made reproducible: all the source code, scores and trained models of this approach are made publicly available.

show abstract

Section: Cuhk Cufs (Vis-sketch)supporting

confidence: 60%

“…An Equal Error Rate of 1.65% was achieved under the Cross-Spectral Iris/Periocular Recognition Competition [30].…”

Section: Feature Learning Based Methodsmentioning

confidence: 96%

Section: Feature Learning Based Methodsmentioning

confidence: 99%

Section: Casia (Vis-nir)mentioning

confidence: 99%

Section: Nivl (Vis-nir)mentioning

confidence: 99%

See 3 more Smart Citations

Heterogeneous Face Recognition Using Domain Specific Units

Pereira

Anjos

Marcel

2019

IEEE Trans.Inform.Forensic Secur.

Self Cite

View full text Add to dashboard Cite

show abstract

Deep Sclera Segmentation and Recognition

Rot

Vitek

Grm

et al. 2019

Handbook of Vascular Biometrics

View full text Add to dashboard Cite

In this chapter, we address the problem of biometric identity recognition from the vasculature of the human sclera. Specifically, we focus on the challenging task of multi-view sclera recognition, where the visible part of the sclera vasculature changes from image to image due to varying gaze (or view) directions. We propose a complete solution for this task built around Convolutional Neural Networks (CNNs) and make several contributions that result in state-of-the-art recognition performance, i.e.: (i) we develop a cascaded CNN assembly that is able to robustly segment the sclera vasculature from the input images regardless of gaze direction, and (ii) we present ScleraNET, a CNN model trained in a multi-task manner (combining losses pertaining to identity and view-direction recognition) that allows for the extraction of discriminative vasculature descriptors that can be used for identity inference. To evaluate the proposed contributions, we also introduce a new dataset of ocular images, called the Sclera Blood Vessels, Periocular and Iris (SBVPI) dataset, which represents one of the few publicly available datasets suitable for research in multi-view sclera segmentation and recognition. The datasets come with a rich P. Rot and M. Vitek are first authors with equal contributions.

show abstract