PSYB51: Lecture 3: Spatial Vision
pinhole plus lens focus light in retina - precondition of vision
achieve neurally -
more than percieving things in 3d - spatial vision
What does visual world look like? or designed for.
Natural scenes that brain will extract and filter it our and individual neurons serve as filters
Diff. version of gratings
what are the bound or limitations of acuity
primary visual pathway - retina - synapses - cortex to area 17
functional prop. of stripe cortex
Natural Scene Statistics: Why care about gratings?
gutenberg's invented printing press - impact on why it is impt. to represent something
gratings in visual world is the building block
e.g. Fourier's transform
assoc. with natural world as well
Printing press metaphor
huge impact in education
Cambridge own 122 books
need to be copied these books by hand so printing press can mass produced books - has a
big impact on education in Europe.
frame and cover with ink cover paper and print
takes some time to create book - but much quicker than doing it by hand
invention not new at all bec. in Korea earlier
why gutenberg invention - is it bec. of bias found out in western countries? culture bias?
reason printing press in Europe is successful bec. the way to represent language much
more efficient when you use it on alphabet.
need this movable types - efficient in english than chinese.
Why care about sine wave gratings?
create filters that specialize and detect 1 building block at a time e.g. only detect A or
Zindividual neurons do it. so visual sys. and brain can decompose it to simpler elements to
have simpler definition that way.
3 types of gratings
1. white and black - luminance profile - gratings show seq. of bright and darker rectangular
shape - clear boundaries sinusoidal grating - pt. smooth transition from dark and bright - sine waves up - bright
down dark -
little bright stripes with darkness
GAbor - 3rd type of grating - instead of black blue - stripe surrounded by darkness or
to get idea of luminance distribution - bell shape function - gaussian shape
contrast - see only inside window with blurred boundaries
3 types used in visual science as early function of visual sys.
3 characteristics of gratings
freq. day and night - happening once every 24 hours
freq. cycles in sound measured in hertz
spatial freq - visual degree - CPD - cycle per degree how many cycle in visual degree
blue sinosoidal - one cycle - one time in bright and 1 dark cycle not repeated yet -
compared with blue and green curve - green twice degree frequencies blue curve
Amplitude - distance from 0 - amplitudes e.g. how loud sound and light - how much
contrast grating or light have
red curve - similar - freq. not change and amplitude but phase change
curve upward blue downward
bec. little pcs. of gratings - see fuzzy edges - small pc. of gratings
decompposng is frst step in visual sys. decompose sine waves helps us to encode visual
world in efficient way in terms of simple mechanism
works for musice - digital waves of coding music - qualitive of music on how it is stored -
plate disks - analog way -
cds do it diff, take sound wave of music and decompose things in sine waves
speakers - operate better for diff. ranges of freq. can have one freq. bec. it is infinitive but
can be broken down to simplifiy and hear better.
in visual sys.transform step by step in process
do it in horizontal and vertical
transform and take out low freq - blurred image emphasizes lowoer freq. iage info. about
transition of black and white that is slow.
if pick up low freq. pattern of zebra lost but if pick up high freq. not see anymore what is
black or white but emphasize picking out boundaries of transitions - get very different
types of info bet. hogh or low freq.e.g. forest - cant distinguish anymore if low freq.
we can distinguish hundreds of pictures in 100 sec.
extracting freq. gives rise to impt. aspect of information
High Spatial freq. Mask
illusion - contradictory info. Lincoln illusion - higher freq. interferes with loss of info.
high and low spatial freq. - spatial freq range pic. of abraham lincoln
low freq. range face of lincoln - if high freq. range just pixels -grate of squares - if stand
near that picture - interfering to percieve face - depends on how close you are to image -
bec. high freq. more or less visible
gratings - arbitrary way on perception ways
best set of building blocks-build language to represent - alphabet more efficient and for
visual images do stat.methods identify components of images
Bell and Sjevnoski
when filter many photographs of natural scenes - get line of inflated graph - every natural
image is composition of this components little lights and pcs of gratings making up all kinds
efiiciencey better thananything else
in pictograms - chinese more redundancy - far more chinese types - more efficient to have
independent components than analogous fines
how is this relevant for biology and psy. at the end - interesting to analyze photos -pictures
approximations of real life not 3d not represent complete visual reality - close enough
simplifications to find out something on first place - something about brain - criteria of
brain to reconstruct something - past experience - experience with natural images - scenes
visual sys. adapt - we dont store every little impression we get but store in efficient way -
store many images not enough space - thoretically have many images in world - shadows
would fall differently - sun shine clouds 0 seasons not fit on brain - not efficient enough
human brain just enough that body burns - energy
across evolution not possible make use of something more efficient
efficient coding - biological relevance –
pictures are simplification of world - natural images have certain systematic features
visual sys. adapt and learn and try to be efficient - plastic
therefore - building blocks can be expected - see it somewhere - ability to pck spatial freq.
mech. that fucntion as sys. pick out building blocks.
system detect individual -
filters -pick out building blocks - detection mechanism specialize for grating see edges and
individual neurons detect lines and edges
perceive things in certain pt. spatial freq. is high not see anymore - narrowness of stripes
are limited -
other limitation is contrast contrast sensitivity or contrast threshold
basic way to test it acuity
Same or different?
pick out smallest detail that can be resolved
fixated x in center
pt. quantify visual acuity cause hard to distinguish depends also on how much sleep you get
Measuring visual acuity
sharp vision - 20/20 acuity is normal
distance in 20 feet normal acuity see in distance of 20 feet
good visual acuity - minimum visual angel of cycle of 0.01 degrees
visual angle can still be resolved
less angle then can't see anymore
can also convert no. in cpd
if you have tiny fraction of 1 visual degree= 60
Opthalmologist measure acuity and vision science
how large receptor fields are individually
cones in the fovea
photoreceptors pts at 1 part centers light in discreet way
take in whatever light shine on entire surface - absolute limit of visual acuity
Acuity is further by convergence (multiple photoreceptors projecting into single bipolar
cell) - see in fovea in retina
convergence not true in fovea
grating- center surround in ganglion cells of receptive fields photoreceptors have receptive
fields - extend of photoreceptors
super impose in gratings
light shines into them and gets stimulated
range that there are no light from retina so not stimulated
all photoreceptors together tell me that there is grating
possibility of seeing grating
size of receptor fields cones is
every little bubble covers grating
cones not matter not care but for rods - black and white side in 2nd stripes image which part of them light shines - average of entire surface of photoreceptors that state
which photorec. get stimulated or not
grating freq. stimulated too high.
some pt not see it as grating anymore
blurred and continues
castle with brick wall - curve on right side but not what is really in there - transition to
being resolve somthing or not resolve at all
visual world is analogue continues but vis. sys not perceive that bec. photorecptors cares
only when light is present
the pt. where we see aliasing is by snellen test
standardize letters with diff. orientation - pt of E to measure acuity perceive E sufficiently