Pre-processing Techniques to Improve the Efficiency of Video
Identification for the Pygmy Bluetongue Lizard
Damian Tohl
1
, Jim S. Jimmy Li
1
and C. Michael Bull
2
1
School of Computer Science, Engineering and Mathematics, Flinders University, South Road, Tonsley, SA, Australia
2
School of Biological Sciences, Engineering and Mathematics, Flinders University, Bedford Park, SA, Australia
Keywords: Video Identification, Pygmy Bluetongue Lizard, Curvature, DWT, SIFT.
Abstract: In the study of the endangered Pygmy Bluetongue Lizard, non-invasive photographic identification is
preferred to the current invasive methods which can be unreliable and cruel. As the lizard is an endangered
species, there are restrictions on its handling. The lizard is also in constant motion and it is therefore
difficult to capture a good still image for identification purposes. Hence video capture is preferred as a
number of images of the lizard at various positions and qualities can be collected in just a few seconds from
which the best image can be selected for identification. With a large number of individual lizards in the
database, matching a video sequence of images against each database image for identification will render
the process very computationally inefficient. Moreover, a large portion of those images are non-identifiable
due to motion and optical blur and different body curvature to the reference database image. In this paper,
we propose a number of pre-processing techniques for pre-selecting the best image out of the video image
sequence for identification. Using our proposed pre-selection techniques, it has been shown that the
computational efficiency can be significantly improved.
1 INTRODUCTION
The Pygmy Bluetongue Lizard is an endangered
species which was thought to be extinct for thirty
years. They are found exclusively in remnant
fragments of native grassland in South Australia’s
mid-north (Li et al, 2009), (Tohl et al, 2013),
(Staugas et al, 2013), (Schofield et al, 2013).
Identification of individual lizards is essential for
ecological studies. One commonly used method is
toe clipping. It is a highly invasive method whereby
digits are removed from the feet of the lizards. The
accuracy of this method can be affected by the fact
that natural toe and foot loss can occur in lizards in
nature (Hudson, 1996). Due to the Pygmy
Bluetongue Lizards endangered status, a non-
invasive identification method, such as photo
identification using the Scale Invariant Feature
Transform (SIFT) method (Lowe, 2004) is preferred.
As the Pygmy Bluetongue Lizard is an
endangered species, there are restrictions on the
amount of time a lizard can be captured for and the
amount of handling. The lizards are captured in the
field and placed in a Perspex box in which the video
is captured and measurements are taken. There is
little control over the lighting conditions and their
posture cannot be easily manipulated as the lizards
are alive and constantly moving. It is therefore
preferred to capture a video which is an image
sequence of the lizard. However, it is very
computationally inefficient to match every image in
the sequence with every image in the database using
SIFT, especially when the database could contain
over hundreds of lizards. A number of pre-
processing techniques are therefore proposed for
pre-selecting the best image out of the image
sequence of the video prior to identification of the
lizard using SIFT.
From our experimental observation, the accuracy
of SIFT identification depends on a number of
factors including the degree of sharpness of the
image and the difference of body curvature from the
reference image in the database. Due to both camera
and lizard movement and the time delay required to
refocus by the camcorder, some images will be non-
identifiable because of motion blur and out of focus.
To determine the degree of sharpness of an image,
the total energy of the high frequency components of
the image is evaluated, based on the fact that sharp
details contain high frequency components. The
623
Tohl D., Li J. and Bull C..
Pre-processing Techniques to Improve the Efficiency of Video Identification for the Pygmy Bluetongue Lizard.
DOI: 10.5220/0005317306230629
In Proceedings of the 10th International Conference on Computer Vision Theory and Applications (VISAPP-2015), pages 623-629
ISBN: 978-989-758-089-5
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)