Transcript Digital Signal Processing

Input technologies
• All require some form of data acquisition
– e.g. Image scanner, Microphone
• Once acquired, if the signal is not already digital,
it will need to be converted using a A to D
converter.
• Image scanners for scanning fingerprints for
example, tend to use one of two scanning
technologies: optical and capacitive.
– What is capacitive?
General recognition block diagram
Input
Electronics
Image
capture
Sound
capture
Pen
capture
A to D
Pre-proc
Reco
Output
Electronics
Monitor
Brail pins
Printer
Speaker
Intro to common components
• Hardware is used at the extreme input and
output so a basic understanding is necessary
• Resistor (Ohms)
– Current directly proportional to Voltage
– Current inversely proportional to resistance
• Capacitor (Farads)
– Charge storage
– Current is proportional to capacitance and rate of
change of voltage
– Capacitance is directly proportional to plate area and
inversely proportional to plate distance
Common components
• Voltage amplifier
– Typically inverts so Vout = -A.Vin
• Charging a capacitor via a resistor
– Capacitor charges non-linearly
– http://www.lon-capa.org/~mmp/kap23/RC/app.htm
Resistor Capacitor
Chargingcurrent: i  Iˆe
DishargingVoltage: v  Vˆe


t
CR
t
CR
Dischargecurrent: i   Iˆe

t
CR
ChargingVoltage: v  Vˆ (1  e

t
CR
)
Amplifiers
R2
R1
-
Vout = -Vin.R2/R1
Vin
+
C
R1
Vin
+
Vout
Vout  
1 t
Vin  d

0
RinC
Capacitive scanning
• Skin has electrical properties such as
resistance and capacitance so can be
used to control electric current
• A touch switch for example relies on either
the resistance of the skin to bridge two
contacts to pass an electric current or to
bridge two contacts to connect a
capacitance in a circuit.
• An array of electrical contacts placed at a
resolution of less than the width of a finger
print ridge
• The nearer the skin is to the contact, the
higher the capacitance value – so a ridge
would create a high capacitance value at
each set of contacts it touches
• Each set of contacts connects the
feedback loop of an integrating amplifier
C
R
i
V
i
n
n
P
+
V
ou
t
• Integrating amplifier. Keep R and t constant so
Vout is dependant on C
Vout
1 t
.

Vin  d

0
RinC
C
Rin
P
Vout
Vin
+
Microphone
• 3 main types
– Dynamic
• Electromagnetic using moving coil
– Condenser
• Electrostatic. Needs a voltage source
– Electret
• Similar to condenser but does not need voltage
source
Dynamic microphone
• Like a loudspeaker
in reverse
Condenser microphone
Area
Capacitance 
Distance
•Sound cards provide voltage through connector
Electret microphone
• The electret condenser mic uses a special
type of capacitor which has a permanent
voltage built in during manufacture. This is
can be likened to a permanent magnet, in
that it doesn't require any external power
for operation. Otherwise it operates in the
same way as a condenser microphone
Capacitance scanner operation
• Finger is placed on the scanner
• All capacitances are short circuited to
discharge any residual charge in the
contacts
• The resistance is constant and a DC
voltage is applied to the input for a fixed
time
• The voltages at the outputs of the
integrators is input to an A to D converter
Loudspeaker
Optical Scanning
• Generally use a Charge Coupled Device
(CCD).
– Array of photodiodes which generate an
electric current in response to being struck by
light photons – i.e. output is analogue.
– An A to D converter converts the analogue
variation from each cell into a binary number
representing brightness
Image Sensor Operation
• Charge Coupled Device
image sensor
• Background is image
sensor zoomed in
• CCD measures brightness
• Tiny lenses direct light onto
filtered photosensitive regions
• More green than red to better
match the eye
Bayer filter
• Eye responds mostly to
green so as many green
filters as red+blue
• Demosaic – interpolate a single
pixel colour by interpolating
nearest neighbours as each
pixel only records one colour so
the actual colour at that point is
the average of it and the
surrounding pixels
CCD Operation
CCD Operation
• Charge is moved down 1 row at a time then
clocked out to an amp and A to D converter
• CCD device used to photograph the finger
when placed on the glass
• Light source is usually an array of LEDs to
enable sharp capture of fingerprint ridges
– Image processing techniques (covered later in
the module) check the contrast and
sharpness of the captured image and adjusts
the light source (or generates instructions to
the user) if necessary
Pros, Cons and Limitations
• Capacitance scanner can not be shown a
picture of a fingerprint for recognition so is not as
easily fooled as CCD
• Capacitance scanners tend to be smaller
• Both can be fooled by latex copy or by chopping
off someone's finger
• Some systems have heat / pulse sensors
– Can put latex copy over real finger to fool these
• Better to use additional biometrics / password
Fingerprint Recognition
• Comparing one whole fingerprint to
another by direct comparison is infeasible
as the contrast may be different, pressure
and orientation may be slightly different …
• Most recognition technologies look for
specific features whether this is fingerprint
or speech recognition as a lot of the data
defining uniqueness is redundant
Fingerprint characteristics
• The scanned image shows ridges as black
and valley’s as white areas
• Characteristics are called minutiae
(mi-NOO-shee-ee )
– Ridge ends (A)
– Bifurcations (B) ridge splits from 1 to 2
• Image needs to be
pre-processed to
enhance it
• Algorithms needed to
used to identify
minutiae
General recognition block diagram
Input
Electronics
Image
capture
Sound
capture
Pen
capture
A to D
Pre-proc
Reco
Output
Electronics
Monitor
Brail pins
Printer
Speaker
Summary
• As will be seen with most recognition
technologies:
– data is captured
– analogue to digital conversion
– enhanced (e.g. contrast enhance)
– Feature extraction (in this case minutiae)
– Recognition algorithm