Dynamic best spectral bands selection for face recognition

Bouchech, Hamdi Jamel; Foufou, Sebti; Abidi, Mongi A.

doi:10.1109/ciss.2014.6814081

Cited by 5 publications

(7 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…8 Another challenge is when data collection is performed using sensors that operate at different spectral bands (VIS, UV, infrared (IR)). 9,10 Differences in appearance between images sensed in the VIS and the UV band are due to the properties of the object being imaged. Ultraviolet radiation is similar to visible light in all the physical aspects, the UV region has wavelengths lower than that of visible light.…”

Section: Introductionmentioning

confidence: 99%

Can we match ultraviolet face images against their visible counterparts?

2015

View full text Add to dashboard Cite

In law enforcement and security applications, the acquisition of face images is critical in producing key trace evidence for the successful identification of potential threats. However, face recognition (FR) for face images captured using different camera sensors, and under variable illumination conditions, and expressions is very challenging. In this paper, we investigate the advantages and limitations of the heterogeneous problem of matching ultra violet (from 100 nm to 400 nm in wavelength) or UV, face images against their visible (VIS) counterparts, when all face images are captured under controlled conditions.The contributions of our work are three-fold; (i) We used a camera sensor designed with the capability to acquire UV images at short-ranges, and generated a dual-band (VIS and UV) database that is composed of multiple, full frontal, face images of 50 subjects. Two sessions were collected that span over the period of 2 months. (ii) For each dataset, we determined which set of face image pre-processing algorithms are more suitable for face matching, and, finally, (iii) we determined which FR algorithm better matches cross-band face images, resulting in high rank-1 identification rates. Experimental results show that our cross spectral matching (the heterogeneous problem, where gallery and probe sets consist of face images acquired in different spectral bands) algorithms achieve sufficient identification performance. However, we also conclude that the problem under study, is very challenging, and it requires further investigation to address real-world law enforcement or military applications. To the best of our knowledge, this is first time in the open literature the problem of cross-spectral matching of UV against VIS band face images is being investigated.

show abstract

Section: Introductionmentioning

confidence: 99%

Can we match ultraviolet face images against their visible counterparts?

2015

View full text Add to dashboard Cite

show abstract

“…Hyperspectral imaging techniques for face recognition have provided promising results in the field of biometrics, overcoming challenges such as pose variations, lighting variations, presentation attacks and facial expression variations [27]. The fusion of narrow-band spectral images in the visible spectrum has been explored to enhance face recognition performance [28]. For example, Chang et al [21] have demonstrated that the fusion of 25 spectral bands can surpass the performance of conventional broad band images for face recognition, mainly in cases where the training and testing images are collected under different types of illumination.…”

Section: Spectral Face Recognitionmentioning

confidence: 99%

Deep Sparse Band Selection for Hyperspectral Face Recognition

Taherkhani

Dawson

Nasrabadi

2020

Hyperspectral Image Analysis

View full text Add to dashboard Cite

Hyperspectral imaging systems collect and process information from specific wavelengths across the electromagnetic spectrum. The fusion of multispectral bands in the visible spectrum has been exploited to improve face recognition performance over all the conventional broad band face images. In this book chapter, we propose a new Convolutional Neural Network (CNN) framework which adopts a structural sparsity learning technique to select the optimal spectral bands to obtain the best face recognition performance over all of the spectral bands. Specifically, in this method, images from all bands are fed to a CNN, and the convolutional filters in the first layer of the CNN are then regularized by employing a group Lasso algorithm to zero out the redundant bands during the training of the network. Contrary to other methods which usually select the useful bands manually or in a greedy fashion, our method selects the optimal spectral bands automatically to achieve the best face recognition performance over all spectral bands. Moreover, experimental results demonstrate that our method outperforms state of the art band selection methods for face recognition on several publicly-available hyperspectral face image datasets.

show abstract

“…Dynamic systems are more accurate since each subject is treated separately, while static systems are less time consuming since the best bands selection is done only one time. From the four studied works in this paper, the work in [11] presented a dynamic system while the three others proposed static systems for best spectral bands selection.…”

Section: Selectionmentioning

confidence: 99%

“…However, BSBS could be dynamic too as proposed in [11]. In this work, the authors assumed the existence of two probability density functions (PDFs) F good and F bad that determine the probability of a given spectral band k to belong to the set of good spectral bands and bad spectral bands respectively.…”

Section: B Dynamic Best Spectral Bands Selectionmentioning

confidence: 99%

“…In this paper, we have reviewed the existing few research works related to the problem of best spectral bands selection for face recognition. The theory behind four contributions; three static [8], [10], [12] and one dynamic [11] has been presented. The performances of these BSBS algorithms to enhance the recognition accuracy of MBLBP, HGPP and LGBPHS algorithms were compared.…”

Section: B Selection Of Best Approaches For Best Spectral Bands Selementioning

confidence: 99%

See 1 more Smart Citation

A comparative study of best spectral bands selection systems for face recognition

Bouchech

Foufou

2014

2014 IEEE/ACS 11th International Conference on Computer Systems and Applications (AICCSA)

Self Cite

View full text Add to dashboard Cite

Multispectral images (MI) have shown promising capabilities to solve problems resulting from high illumination variation in face recognition. However, the use of MI, with the huge number of captured spectral bands for each subject, is impractical unless a system for best spectral bands selection (BSBS) is used. In this work, first we give an up to date overview of the existing BSBS techniques proposed for face recognition. We aim to highlight the imporatnce of this component of MI based systems. The reviewed techniques are then experimented using the multispectral face database IRIS − M 3 to compare their performances. To the best of our knowledge this is the first study that reviews and compares existing techniques for BSBS. The Obtained results emphasized the importance of setting up techniques for BSBS with MI based systems as well as the need for new techniques that investigate better the increasing speed of image processing tools.

show abstract

Dynamic best spectral bands selection for face recognition

Cited by 5 publications

References 15 publications

Can we match ultraviolet face images against their visible counterparts?

Can we match ultraviolet face images against their visible counterparts?

Deep Sparse Band Selection for Hyperspectral Face Recognition

A comparative study of best spectral bands selection systems for face recognition

Contact Info

Product

Resources

About