Do you recognise me?
Julia Farrugia
A
utomatic facial recognition
could change the world of
Sketches and photos have different
natures (modalities)—photos are
law enforcement. Profile photos
generally captured using a digital
of suspects are rarely available, so
camera, while sketches may be hand-
investigators still rely on face sketches
drawn or computer generated. In
based on eyewitness descriptions.
order to tackle this problem, Farrugia
Julia Farrugia (supervised by Dr Ing.
developed an inter-modal approach
Reuben Farrugia) implemented an
to sketch retrieval. Without changing
automatic face recogniser that is able
the nature (modality) of the images,
to retrieve a photo based on a sketch.
common features in the sketch and
This narrows down the number of
photo were used as a basis for retrieval.
potential criminals before trails start
Testing was carried out using the
to go cold.
Chinese University of Hong Kong
The future of transport
Brandon Spiteri
T
he world has globalised. People
friction since the vehicle floats on
and cargo need to get about in
electromagnetic waves that make
cheaper, faster ways that use better
transport technologies. Magnetic
levitation is one way to achieve
higher speeds at a cheaper
fuel cost whilst offering
a smoother ride.
There is less
this transport method very efficient.
Brandon Spiteri (supervised by Dr
Ing. Maurice Apap and Prof. Joseph
Cilia) designed and built a model
in which a vehicle was moved at
constant speed whilst levitating 1 cm
above the track. Spiteri identified
three levitation techniques. Firstly,
the German approach eliminates
needing wheels to initially move the
train, but requires complex control
methods. Secondly, the Japanese
approach requires wheels to
initially move the train, but
achieves higher speeds than
German technology. Lastly,
the MDS type system is
still being developed
Students
but aims for higher
14
The Japanese MLX01 Maglev train
Photos used with permission: X. Wang and X. Tang,
“Face Photo-Sketch Synthesis and Recognition,”
IEEE Transactions on Pattern Analysis and Machine
Intelligence (PAMI), Vol. 31, 2009
(CUHK) student database, which
match between sketch and photo. To
filtering the photos according to
contains 188 photo-sketch pairs.
improve these results, texture features
gender and by experimenting on larger
The implementation makes use of
of the query sketch and each photo in
datasets with subjects from different
an Active Orientation Model (AOM),
the dataset were extracted using Local
ethnicities, wearing glasses, or having
which is freely available. 68 strategic
Binary Patterns (LBP). The distance
facial hair. Advancements in computer
points on a query sketch and suspect
was again calculated but included the
vision means that soon humans will
photo are plotted. Dots depict features
texture features. The results were then
not be the only eyes narrowing down
like eyebrows, hairline, and nose. The
merged with the distances obtained
possible suspects.
distance was calculated between the
using the AOM method. Giving a higher
respective points on the sketch and
priority to the distances obtained using
This research was carried out as part
photo. The smaller the difference
the texture features increased the
of a Bachelor of Science in Computer
in distances, the closer the match.
recognition rate to 60.11%.
Engineering at the Faculty of ICT,
55.85% of tests resulted in a correct
Results could be improved by
University of Malta.
The model built by Spiteri, levitating over the track.
speeds than the German model
The strength and polarity of the
a less polluting and more efficient
without the need for wheels.
electromagnet varies with the size
system. Fresh graduates Justin Zarb
and direction of the electrical current
and Luke Lapira recently proposed
model by using an industrial-power
passed through it. By manipulating
a plan called Maltarail (elevated,
DC motor. Levitation was achieved by
the electromagnets the vehicle moved
suspended trains running on a single
using magnets of similar polarity that
forward. The built model achieved a
rail) to government. This project has
repel each other. Opposing permanent
top speed of 1.41 km/hr. In his study,
been submitted to the European
magnets were installed on the track
Spiteri proved the energy efficiency
Investment Initiative. Such a project
and vehicle. Permanent magnets
of these systems: the model uses
would place Malta on par with
retain their magnetic properties
the same energy as a 12 W bulb,
European transport leaders.
(North and South poles) even when
much less than a train on wheels.
no current or electromagnetic field is
Magnetic levitation will shape the
This research was carried out as
present. The train moved by having
future of transportation worldwide.
part of a Bachelor of Engineering
permanent magnets on the track
Monorail may be vital to reduce
at the Faculty of Engineering,
and electromagnets on the vehicle.
Malta’s transport problems to have
University of Malta. Students
After this research, Spiteri built a
15