cap5415'computer/vision/...
Post on 09-Mar-2020
4 Views
Preview:
TRANSCRIPT
CAP5415'Computer/Vision/Lecture/1'Introduc8on//
Ulas/Bagci/bagci@ucf.edu/
8/25/15/Lecture/1/'/Introduc8on/
1/
About/Me/(Short/Bio)/
• 2003:%Graduated/from/EE/Bilkent/University/(Turkey)./• 2005:%MSc/EECS/Koc/University/(Turkey)./• 2006(2010:%Marie/Curie/Fellow,/University/of/NoUngham/(UK)/and/
University/of/Pennsylvania/(USA),/Computer/Science/and/Radiology/Departments.//
• 2010(2012:%ISTP/Research/Fellow,/Radiology,/NIH./• 2012(2013:%Senior/Research/Fellow,/Radiology,/NIH./• 2013(now:%Staff/scien8st/and/Lab/Manager/(CIDI),/Radiology,/NIH./• 2013(2014:%Leading/image/analysis/scien8st/of/Bio'terrorism/bio'defense/
studies/in/the/USA/(NIH,/Integrated/Research/Facility)./• 2015(Now:%Assist./Prof./at/CRCV,/UCF/
• Research%Interests:%Medical/computer/vision,/image/processing/and/analysis,/biomedical/and/clinical/imaging/applica8ons,/paZern/analysis,/sta8s8cal/machine/learning./
/
8/25/15/Lecture/1/'/Introduc8on/
2/
Course/Syllabus//• Class/8me:/Tuesday/Thursday/3pm'4.15pm/• Office/hours:/HEC221/4.30pm'6pm/
• 5/Programming/assignments/(PA)/50%/(each/10%)//– First/programming/assignment/(PA'0)/PA/will/be/op8onal,/it/will/be/graded,//and/counted/as/
BONUS.///
• 2/mini/projects/50%/(each/25%)/• Python/is/the/required/language/for/PA./
– Any/IDE/is/fine,/CANOPY/is/free/to/use./
• Processing,/Matlab,/C/C++,/Java,/Python,../can/be/used/for/mini/projects./
• No/required/books,/but/op8onal/ones:/– Szeliski,/Computer/Vision:/Algorithms/and/Applica8ons,/Springer/2010/(online/drai)/– Shah,/Fundamentals/of/Computer/Vision/(available/from/the/course/webpage)/– Python/for/Computer/Vision/(available/online)/
8/25/15/Lecture/1/'/Introduc8on/
3/
Course/Goals/
• Introductory/level/computer/vision/course,/suitable/for/graduate/students./– Image/filtering,/edge/detec8on/– Mo8on/and/op8cal/flow/– Region/Shape/Segmenta8on/– Shape/modeling/and/analysis/– Deep/Learning/for/Computer/Vision/– Imaging/Geometry,/Camera/Modeling,/Calibra8on/
8/25/15/Lecture/1/'/Introduc8on/
4/
Computer/Vision/8/25/15/Lecture/1/'/Introduc8on/
5/Applica8ons/
Ex./Object/Recogni8on/
• Problem:/Given/an/image/I,/does/I/contain/an/image/of/a/person?/
8/25/15/Lecture/1/'/Introduc8on/
6/
Ex./Object/Recogni8on/
• Problem:/Given/an/image/I,/does/I/contain/an/image/of/a/person?/
8/25/15/Lecture/1/'/Introduc8on/
7/
YES/
Ex./Object/Recogni8on/
• Problem:/Given/an/image/I,/does/I/contain/an/image/of/a/person?/
8/25/15/Lecture/1/'/Introduc8on/
8/
NO/
Ex./Object/localiza8on/8/25/15/Lecture/1/'/Introduc8on/
9/
Ex./Human/Detec8on/8/25/15/Lecture/1/'/Introduc8on/
10/
Ex./Face/Recogni8on/8/25/15/Lecture/1/'/Introduc8on/
11/
Ex./Image/Search/8/25/15/Lecture/1/'/Introduc8on/
12/
Tons/of/applica8ons,/Big/market,/…/8/25/15/Lecture/1/'/Introduc8on/
13/
/350 million photos uploaded daily
100 hours of movies uploaded per hour
60,000 movies
Open'Universe/Face/Iden8fica8on/8/25/15/Lecture/1/'/Introduc8on/
14/
Barack Obama
Bob
Alice
News Article: Label Important Figures
Social Network: Tag Facebook Friends
Open'Universe/Face/Iden8fica8on/8/25/15/Lecture/1/'/Introduc8on/
15/
Find Angelina Jolie and George Clooney
Open'Universe/Face/Iden8fica8on/8/25/15/Lecture/1/'/Introduc8on/
Find Angelina Jolie and George Clooney
Ex./Facial/expression/8/25/15/Lecture/1/'/Introduc8on/
17/
Ex./Fa8gue/detec8on/8/25/15/Lecture/1/'/Introduc8on/
18/
Ex./Lip'reading/8/25/15/Lecture/1/'/Introduc8on/
19/
Video/Surveillance/and/Monitoring/8/25/15/Lecture/1/'/Introduc8on/
20/
Object/detec8on/ Object/tracking/Object/categoriza8on//and/classifica8on/
Event/or/Ac8vi8es/Recogni8on/
UAVs:/Unmanned/Aerial/Vehicles/(drones)/
8/25/15/Lecture/1/'/Introduc8on/
21/
Global/Hawk/
Predator/
Microdrone/
Ex./Detec8on/in/Videos/8/25/15/Lecture/1/'/Introduc8on/
22/
Ex./Object/Tracking/8/25/15/Lecture/1/'/Introduc8on/
23/
Ex./Wide/Area/Surveillance/8/25/15/Lecture/1/'/Introduc8on/
24/
Ex./Tracking/(mul8'object)/8/25/15/Lecture/1/'/Introduc8on/
25/
Ex./Human/Ac8on/Recogni8on/8/25/15/Lecture/1/'/Introduc8on/
26/
Bench Swing Dive Swing Run
Kick Lift Ride Golf Swing Skate
9"actions,"142"videos"(UCF"database,"Shah)."
Ex./UCF/YouTube/Ac8on/Dataset/8/25/15/Lecture/1/'/Introduc8on/
27/
Cycling( Diving( Golf(Swinging( Riding(
Juggling( Basketball(Shooting( Swinging( Tennis(Swinging(
Volleyball(Spiking( Trampoline(Jumping( Walking((Dog(
Ex./High/Density/Crowded/Scenes/8/25/15/Lecture/1/'/Introduc8on/
28/
Political Rallies Religious Festivals Marathons High Density Moving Objects
Ex./Coun8ng/in/Extremely/Dense/Crowd/Images/
8/25/15/Lecture/1/'/Introduc8on/
29/Proposed/Method/by/Idrees/and/Shah=640/Ground/truth=634/
8/25/15/Lecture/1/'/Introduc8on/
30//////Proposed/Method=1468/Ground/truth=1428////////////
8/25/15/Lecture/1/'/Introduc8on/
31///////Proposed/Method=2496/Ground/truth=2319/
Ex./Visual/Business/Recogni8on/8/25/15/Lecture/1/'/Introduc8on/
32/
Ex./Medical/Computer/Vision/8/25/15/Lecture/1/'/Introduc8on/
33/
Ex./X'Ray/Imaging/Radiography/8/25/15/Lecture/1/'/Introduc8on/
34/
The/first/published/medical/image//was/a/radiograph/of/the/hand/of//Wilhelm/Conrad/Roentgen’s/wife/in/1895.//!Nobel!Prize!in!Physics!1901./
Ex./Ultrasound/imaging/and/Analysis/
8/25/15/Lecture/1/'/Introduc8on/
35/
Ex./Renal/Artery/Blood/Flow/Es8ma8on/by/Computer/Vision/Techniques/
8/25/15/Lecture/1/'/Introduc8on/
36/
stenosis/is/seen/eca:/external/caro8d/artery/cca:/common/caro8d/artery/ica:/internal/caro8d/artery/
CV/methods/can/help/calcula8ng/All/blood/flow/and/iden8fy/Automa8cally/the/abnormal/regions./
Ex./Computer/Vision/for/Graphics/8/25/15/Lecture/1/'/Introduc8on/
37/
Computed/Tomography////Auto'detec8on/of/tumors/
8/25/15/Lecture/1/'/Introduc8on/
38/
Computer/vs./Human/Vision?/
• In/which/task/computers/are/superior/to/humans?/
• In/which/task/humans/are/superior/to/humans?/
8/25/15/Lecture/1/'/Introduc8on/
39/
8/25/15/Lecture/1/'/Introduc8on/
40/
Anything/interes8ng/in/this/scan?/8/25/15/Lecture/1/'/Introduc8on/
41/
Now?/8/25/15/Lecture/1/'/Introduc8on/
42/
Ex./DWI/and/DTI/8/25/15/Lecture/1/'/Introduc8on/
43/
Tracking/water/oxygen/molecules/characterize/tumors,/and/provide/connec8vity/Informa8on/on/neural/tracks/
8/25/15/Lecture/1/'/Introduc8on/
44/
What/can/you/see/in/this/picture?/8/25/15/Lecture/1/'/Introduc8on/
45/
Now/can/you/see/what/the/picture/is?/
8/25/15/Lecture/1/'/Introduc8on/
46/Credit:!Thompson,!Basic!Vision,!Oxford!Press,!2012.!
Do/they/look/the/same?/8/25/15/Lecture/1/'/Introduc8on/
47/
Visual/Percep8on/
• How/do/we/know/that/the/objects/that/we/see/are/for?/
/
8/25/15/Lecture/1/'/Introduc8on/
48/
Visual/Percep8on/
• How/do/we/know/that/the/objects/that/we/see/are/for?/
• Can/people/“see”/without/being/aware/of/what/they/see?/
/
8/25/15/Lecture/1/'/Introduc8on/
49/
Visual/Percep8on/
• How/do/we/know/that/the/objects/that/we/see/are/for?/
• Can/people/“see”/without/being/aware/of/what/they/see?/
• Why/do/objects/appear/colored?//
8/25/15/Lecture/1/'/Introduc8on/
50/
Visual/Percep8on/• Defini8on:/Process!of!acquiring!knowledge!about!environmental!objects!
and!events!by!extracJng!informaJon!from!the!light!they!emit!or!reflect![Palmer,!2012].!
8/25/15/Lecture/1/'/Introduc8on/
51/
Vision/ Cogni8ve/Ac8vity/
Acquisi8on/of/knowledge/
Visual/Percep8on/• Defini8on:/Process!of!acquiring!knowledge!about!environmental!objects!
and!events!by!extracJng!informaJon!from!the!light!they!emit!or!reflect![Palmer,!2012].!
8/25/15/Lecture/1/'/Introduc8on/
52/
Vision/ Cogni8ve/Ac8vity/
Acquisi8on/of/knowledge/
Percep8on/is/analogous/to/taking/a/picture!/(credit:/Palmer,/2012)/
Visual/Percep8on/8/25/15/Lecture/1/'/Introduc8on/
53/
Vision!is!a!process!in!which!temporally!changing!intensity!and!color!values!in!the!image!plane!!have!to!be!interpreted!as!processes!in!the!real!world!that!happen!in!3D!space!over!Jme!
Vision/vs./Computer/Vision/?/8/25/15/Lecture/1/'/Introduc8on/
54/
Sensing/device/ Interpre8ng/device/ Interpreta8ons/
Picture/Man/Thrash/Bulb/Light/…/
Vision'Eye'Light/8/25/15/Lecture/1/'/Introduc8on/
55/Credit:!Schwarts,!2009.!
Re8nal/Processing://Photoreceptors/absorb/Light/quanta/and/convert/This/radiant/energy/into/Electrical/ac8vity.//They/synapse/on/bipolar/cells,/which/in/turn,/can/s8mulate/ganglion/cells,/thereby/sending/ac8on/poten8als//along/the/op8c/nerve/to/the/LGN(lateral/Geniculte/nucleus)./
Vision/and/Image/Understanding/
• Visual/tasks:/We/use/vision/to/interact/with/environments/and/survive/–/to/navigate/and/avoid/obstacles,/to/recognize/and/pick/up/objects,/to/iden8fy/food/and/danger,/friends/and/enemies,/…/
/
8/25/15/Lecture/1/'/Introduc8on/
56/
Goal/of/Computer/Vision?/• To/bridge/the/gap/between/image/pixels/and/“meaning”/(seman8c)!/
8/25/15/Lecture/1/'/Introduc8on/
57/
What/we/see!/What/computer/sees!/
What/is/a/(digital)/Image?/
• Defini8on:/A/digital/image/is/defined/by/integraJng!and/sampling!con8nuous/(analog)/data/in/a/spa8al/domain/[KleZe,/2014]./
8/25/15/Lecture/1/'/Introduc8on/
58/
LeR!hand!coordinate!system!
Picture/Elements/'/PIXEL/8/25/15/Lecture/1/'/Introduc8on/
59/
PIXELS/are/ATOMIC/ELEMENTS/of/an/image./In/late/1960s,/terminology/‘pixel’/was/introduced/by/a/group/of/scien8st/at/JPL/in/California!/
Image/Types:/Scalar/and/Binary/
• A/scalar/image/has/integer/values///a:/level/(bit)//Ex./If/8/bit/(a=8),/image/spans/from/0/to/255///////////////////////////////////////0/black///////////////////////////////////255/white/Ex./If/1/bit/(a=1),/it/is/binary/image,/0/and/1/only./
8/25/15/Lecture/1/'/Introduc8on/
60/
u 2 {0, 1, ..., 2a � 1}
Image/Type:/RGB/(red,/green,/blue)/
• Image/has/three/channels/(bands),/each/channel/spans/a'bit/values./
8/25/15/Lecture/1/'/Introduc8on/
61/
Image/Type:/RGB/(red,/green,/blue)/
• Image/has/three/channels/(bands),/each/channel/spans/a'bit/values./
8/25/15/Lecture/1/'/Introduc8on/
62/
Image/Format/
• Some/formats:/TIF,/PGM,/PBM,/GIF,/JPEG,/PNG,/RAW,…/
• Medical/Images:/DICOM,/Analyze,/NIFTI,…//
8/25/15/Lecture/1/'/Introduc8on/
63/
HEADER://contains/image/informa8on,/image/size,/pixel/size,/…//DATA:/ /integer,/double,/float,/unsigned/integer,/char,…/
Prac8ce:/Image/Format/Read/Show/
8/25/15/Lecture/1/'/Introduc8on/
64/
/PIL:/Python/Imaging/Library//from/PIL/import/Image/Img/=/Image.open(‘empire.jpg’)////Matplotlib/is/a/good/graphics/library/with/much/More/powerful/features/than/the//PloUng/available/in/PIL/
Color/• If/there/is/no/light,/there/is/no/color!/• Human/vision/can/only/discriminate/a/few/dozens/of/grey/levels/on/a/screen,/but/hundreds/of/thousands/of/different/colors./
– RED/'>//~625/to/780/nm///////////////[long/wavelength]/– ORANGE/'>/~/590/to/625/nm//////[long/wavelength]/– YELLOW/'>/~565/to/590/nm////////[middle/range/wavelength]/– GREEN/'>/~/500/to/565/nm/////////[middle/range/wavelength]/– CYAN/'>/~485/to/500/nm////////////[/middle/range/wavelength]/– BLUE/'>/~440/to/485/nm/////////////[short/wavelength]/– VIOLET/'>/~330/to/440/nm//////////[very/short/wavelength]/
8/25/15/Lecture/1/'/Introduc8on/
65/
COLOR/
• Color/vision/has/evolved/over/millions/of/years./
8/25/15/Lecture/1/'/Introduc8on/
66/
Re8na/of/Human/Eye/8/25/15/Lecture/1/'/Introduc8on/
67/Credit:!KleTe,!2012.!
There/are/three/different/types/of/color'sensi8ve/cones/corresponding/to/(roughly)/RED/(64%/of/the/cones),/GREEN/(about/32%),/and/BLUE/(about/2%)./
6'7/million/cones/120/million/rods/
LIGHT/COLOR'Isaac/Newton’s/House/
8/25/15/Lecture/1/'/Introduc8on/
68/
Video/Clip/
• Sequences/of/frames/• 30/frames/per/second/
8/25/15/Lecture/1/'/Introduc8on/
69/
8/25/15/Lecture/1/'/Introduc8on/
70/
Sequences/of/Images/
8/25/15/Lecture/1/'/Introduc8on/
71/
Available%online!%
Programming/Assignments/8/25/15/Lecture/1/'/Introduc8on/
72/
Thu/Sep/3,/2015 /Programming/Assignment/#0/(BONUS)//Thu/Sep/10,/2015 /Programming/Assignment/#1//Thu/Sep/24,/2015 /Programming/Assignment/#2///Thu/Oct/8,/2015 /Programming/Assignment/#3//Thu/Oct/22,/2015 /Programming/Assignment/#4//Thu/Nov/5,/2015 /Programming/Assignment/#5///Mon/Nov/23,/2015 /Mini'project/#1 //Thu/Dec/10,/2015 /Mini'project/#2 ////Submissions:/ONLINE/(web'course)./Descrip8ve/explana8ons/should/be/included/in/the/code/(i.e.,/comments)./Individual/efforts/are/sought/(collabora8ons/are/allowed/at/discussion/level/only)./Mini'projects/will/be/selected/from/a/list/of/projects.//
top related