Comparison between Thermal and Visible Facial Features on a Verification Approach - Biomedical Engineering Systems and Technologies - page 275

various difficulties. What will occur if the individual's haircut is changed? Is make-up
a determining factor in the process of verification? Would it distort significantly facial
features?
The use of thermal cameras originally conceived for military purpose has expanded
to other fields of application such as control process in production lines, detec-
tion/monitoring of fire or applications of security and Anti-terrorism. Therefore, we
consider its use in human identification tasks in scenarios where the lack of light re-
stricts the operation of conventional cameras. Different looks of the main role from
the film The Saint are shown in figure 1.
Fig. 1. Facial changes of the character played by Val Kilmer in the film The Saint
Val Kilmer modifies his look in this film spectacularly in order to not to be recog-
nized by the enemy.
A correct matching between the test face and that stored in the image database is
expected, although it may seem a hard problem to solve even if natural distortion
effects such as illumination changes or interference are not considered. The recogni-
tion problem should be split in some stages, that is, acquisition of facial images for
testing, features extraction from specific facial regions and finally, verification of the
individual's identity [1].
Currently, computational face analysis is a very lively research field, in which we
observe that new interesting possibilities are being studied. For example, we can
quote an approach for improving system performance when working with low resolu-
tion images (LR) and decreasing computational load.
In [2], it is presented a facial recognition system, which works with LR images us-
ing nonlinear mappings to infer coherent features that favor higher recognition of the
nearest neighbour (NN) classifiers for recognition of single LR face image. It is also
interesting to cite the approach of [3], in which a multi-resolution feature extraction
algorithm for face recognition based on two-dimensional discrete wavelet trans-
form (2D-DWT) is proposed. It exploits local spatial variations in a face image
effectively obtaining outstanding results with 2 different databases.
The images of subjects are often taken in different poses or with different modali-
ties, such as thermographic images, presenting different stages of difficulty in their
identification.
In [4], results on the use of thermal infrared and visible imagery for face recogni-
tion in operational scenarios are presented. These results show that thermal face rec-
ognition performance is stable over multiple sessions in outdoor scenarios, and that
fusion of modalities increases performance.
In the same year 2004, L. Jiang proposed in [5] an automated thermal imaging system
that is able to discriminate frontal from non-frontal face views with the assumption that at
any one time, there is only 1 person in the field of view of the camera and no other