Computer Vision - Gunadarma...

Computer Vision Penginderaan Visual untuk berbagai

keperluan

Dr. Mohammad Iqbal @ 2016

Disampaikan pada seminar nasional “Perkembangan Computer

Vision dan Multimedia" yang dilaksanakan oleh Himpunan

Mahasiswa Teknik Informatika Universitas Nasional pada hari

Rabu, 20 Januari 2016, di Aula Universitas Nasional Blok I lantai 4

Gunadarma University S3, S2, S1 and Proffessional Program

Faculties

1. Computer Science and Information Technology

2. Industrial Technology

3. Economic

4. Civil Engineering and Plan

5. Psikology

6. Literature

Research Organizations

Research Organization University and for every Faculty

Special Science Group Discussion

Lecturer Group Research: Foshema & Scimed

Pusat Studi :

Mikroelektronika dan Pengolahan citra – imaging system dan smart sensor

Robotika dan Multimedia Sistem Multimedia dan Robotik – Implementasi robotic vision dan data set collection

Informatika Kedokteran – Implementasi vision di bidang kedokteran dan kesehatan

Interaksi Manusia dan Teknologi – Evaluasi Interaksi mesin dengan manusia

3

Penelitian Computer Vision di Gunadarma

Menu Seminar kita hari ini…

Penggunaan Vision Hari Ini

Computer Vision Anatomy

Penglihatan (Vision) itu Tidak Sederhana

Apakah Computer Vision?

Kesimpulan

Diskusi

Mengapa perlu belajar tentang Computer vision?

Jutaan citra di capture setiap waktu

Ada jutaan aplikasi yang bisa dibuat berdasarkan CV

http://www.picsearch.com/



• Defenisi Komputer Grafik ? (transformasi 3D->2D) • Defenisi Komputer grafik ? (Modeling vs. Rendering) • Jadi Defenisi Komputer vision (2D->3D) • Defenisi Computer Vision :

• Irisan antara Computer Vision dan Computer Graphics • Menurut para ahli • Permodelan berbasiskan Citra (Image-Based Modeling)

• Disiplin ilmu yang terkait • Kecerdasan Buatan • Dasar Matematika yang dibutuhkan • Kaitan ilmu modern terkini untuk Computer Vision • Lingkup Kurikulum Computer Vision di Universitas


• Kebalikan dari Komputer Grafik

• Pemahaman komputer terhadap Citra (Image Understanding) secara AI, atau menganalisis perilaku (behavior) / pola Citra

• Sensor untuk robotika

• Emulasi Komputer dari penglihatan manusia

Computer vision

World model

Computer graphics

World model

Grafik

Defenisi Komputer Grafik ? (transformasi 3D->2D)

3D geometri

Sifat fisik

Simulasi

proyeksi

Modeling Create model Apply material ke model Tempatkan model di scene Tempatkan light di scene Tempatkan camera

Defenisi Komputer grafik ? (Modeling vs. Rendering)

Directional Light Ambient Light

Point Light

Spot Light

Penggabungan pencahayaan oleh Patrick Doran (2009)

Rendering

Ambil “citra” dengan camera

Dua-duanya dapat selesai dengan commercial software: Autodesk MayaTM ,3D Studio MaxTM, BlenderTM, etc.

9 ILMU LANJUT : Grafik Komputer

Jadi Defenisi Komputer vision (2D->3D)

3D Geometri

Sifat fisik

Estimasi

Defenisi Computer Vision : Irisan antara Computer Vision

dan Computer Graphics

modeling - shape - light - motion - optics - images IP

animation

rendering

user-interfaces

surface design

Computer Graphics

shape estimation

motion estimation

recognition

2D modeling

modeling - shape - light - motion - optics - images IP

Computer Vision

Defenisi Computer Vision [Trucco&Verri’98]

Trucco and Verri: computing properties of the 3D world from one or more digital images

Sockman and Shapiro: To make useful decisions about real physical objects and scenes based on sensed images

Ballard and Brown: The construction of explicit, meaningful description of physical objects from images

Forsyth and Ponce: Extracting descriptions of the world from pictures or sequences of pictures

Defenisi Computer Vision : Permodelan berbasiskan

Citra (Image-Based Modeling)

Images (2D) Geometry (3D)

shape Photometry appearance

+

graphics

vision

image processing

2.1 Geometric image formation

2.2 Photometric image formation

3 Image processing

4 Feature extraction 5 Camera calibration

6 Structure from motion

7 Image alignment

8 Mosaics

9 Stereo correspondence

11 Model-based reconstruction

12 Photometric recovery

14 Image-based rendering

Disiplin Ilmu yang Terkait : Kecerdasan Buatan

Kaitan ilmu modern terkini untuk Computer Vision

Lingkup Kurikulum Computer Vision

Pattern

Recognition

Computer

Vision

Machine

Learning

Multi-view

Geometry

Intelligent

Robotics

Autonomous

Robotics

Multi-Robot

Systems

Image

Processing

Computer

Graphics

Computational

Perception



• Karakteristik Human Vision • Ilusi Adelson Checkerboard • Warna yang konstan (Color Constancy) • Ukuran yang Konstan (Size Constancy) • Ilusi Thatcher

• Area Fokus Komputer Grafik dan Vision –

Hardware & Interaction

• Timeline Teknologi Computer Vision


Mata Manusia Vs Kamera

Penglihatan itu Tidak Sederhana

Penglihatan (vision) prestasi terbesar dari kecerdasan alami (natural intelligence ) manusia

Visual cortex menempati sekitar 50% dari bagian otak Macaque

Seakan2 otak manusia dikhususkan utk menangani urusan vision

Itu raja atau perdana

menteri ya ?

22

Karakteristik Human Vision

Penglihatan adalah proses kontruktif Persepsi kesadaran dari yang kita lihat adalah ILUSI yang

dibuat oleh otak kita (dengan proses yang luar biasa rumit).

Contoh : kecerahan (brightness), warna (color), dan ukuran yang konstan (size constancy)

23

Ilusi Adelson Checkerboard Persepsi brightness adalah fungsi rumit dari nilai piksel

(Image courtesy of Ted Adelson) Brightness constancy problem

Warna yang konstan (Color Constancy)

Warna Piksel sangat dipengaruhi oleh iluminasi

Persepsi dari konstannya suatu warna dikelola oleh otak kita

Sunlight Fluorescent light

(Images courtesy of David Heeger)

Ukuran yang Konstan (Size Constancy) Ukuran obyek VS kedalaman obyek

(Images copyright John H. Kranz, 1999)

Karakteristik Human Vision

Penglihatan akan menyelesaikan tugas tertentu saja dalam konteks yang juga spesifik Umumnya kemampuan visual itu terikat langsung dengan

kebutuhan dan konteks seseorang (kebiasaan hidup, emosional, dll).

Contoh : Thatcher illusion

Ilusi Thatcher

(Due to P. Thompson)

Ilusi Thatcher

Face processing sensitif pada orientasi citranya

HIGH RESOLUTION HIGH BRIGHTNESS LARGE VIEWING ANGLE HIGH WRITING SPEEDS LARGE COLOUR GAMUT HIGH CONTRAST LESS WEIGHT AND SIZE LOW POWER CONSUMPTION LOW COST

Area Fokus Komputer Grafik dan Vision – Hardware &

Interaction

Teknologi Display

Screenless / Hologram technology

Teknologi Surface / Touch screen

Wearable Teknologi

Stereoscopic

Perangkat Input Mouse, tablet & stylus, multi-touch, force feedback, dan game controller lainnya

(seperti Wii), scanner, digital camera (images, computer vision), dsb.

Semua bagian tubuh menjadi devais interaksi: http://www.xbox.com/kinect


Interaction

30

http://www.xbox.com/kinect

Apple iPhone™

Multi form Output Cell Phones/PDAs (smartphones),

laptop/desktops/tablets,

Microsoft PPI display

3D immersive virtual reality systems such as Brown’s new Cave being built at 180 George Street


Interaction

Brown’s old Cave & new Cave

Samsung Galaxy SIII (Android)

Microsoft Surface

Microsoft PPI display

ILMU LANJUT : Interaksi Manusia Komputer

Timeline Teknologi Computer Vision

# Computer Vision History graph from the book of Richard Szeliski



• Langkah2 dalam Pengolahan Citra Digital • Sistem Pencahayaan (Lighting system) • Staging • Lensa dan Kamera • Aplikasi Perangkat Lunak Vision

34


Pada dasarnya sistem Computer atau Machine Vision dibuat untuk

membantu menggantikan keahlian manusia pada bagian visual

Computer Vision Anatomy : Langkah2 dalam Pengolahan

Citra Digital - Level Pengolahan citra

Level 0: Representasi citra (akuisisi, sampling, kuantisasi, kompresi)

Level 1: transformasi Image-to-image (enhancement, restoration, segmentation)

Level 2: Transformasi Image-to-parameter (feature selection)

Level 3: transformasi Parameter-to-decision (recognition and interpretation)


Citra Digital - Kedudukan DIP, ComVis

Image Processing: Levels 0 and 1

Image Analysis: Levels 1 and 2

Computer/Robot Vision: Levels 2 and 3

Computer Graphics/Animation ?

Pendekatan dalam “creating images” atau membuat “visual effects” dari deksripsi yang diberikan pada level sebelumnya.


Citra Digital - Problem Domain

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Image Aquisition

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Image Enhancement

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Image Restoration

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Morphological Processing

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Segmentation

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Object Recognition

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Representation & Description

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Image Compression

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


Citra Digital - Colour Image Processing

Image

Acquisition

Image

Restoration

Morphologic

al Processing

Segmentation

Representation

& Description

Image

Enhancement

Object

Recognition

Problem Domain

Colour Image

Processing

Image

Compression


1. Lighting 2. Staging 3. Lenses 4. Cameras

Computer Vision Anatomy : Lighting

49

Computer Vision Anatomy : Staging

Parameter-parameter penting dalam sistem pencitraan (imaging system).

50

Computer Vision Anatomy : Kamera dan Lensa

Kamera dan Lensa :

Jenis Sensor : CCD Vs CMOS (complimentary metal-oxide semiconductor)

Ukuran Sensor :

Cara Pembacaan : area scanning and line scanning.

CCD/CMOS Size. (Image copyright of Edmund Optics).


Sistem Lensa :

Wide area lens (catadioptric, fisheye) Vs Basic Lens (zoom, macro, telesentric)

Sistem Filter Lensa : Polarization, IR, UV, …

Relationship between focal length, object and camera plane.

(Image copyright of Edmund Industrial Optics).

52


Resolution :

Focus :

Resolusi citra B lebih baik dari A.

(Image copyright of Edmund

Industrial Optic).

Computer Vision Anatomy : Kamera dan Lensa – Model

dan Geometri Kamera

Pinhole camera

Geometric transformations in 2D and 3D

or

53

Computer Vision Anatomy : Kamera dan Lensa – Camera

Calibration

Know 2D/3D correspondences, compute projection matrix

also radial distortion (non-linear)

Aplikasi Perangkat Lunak Vision

HALCON dari MVTEC http://www.mvtec.com/halcon/

HALCON is the comprehensive standard software with an integrated

development environment (IDE) for machine vision that is used worldwide. It leads to cost savings and improved time to market: HALCON's flexible architecture facilitates rapid development of machine vision, medical imaging, and image analysis applications. HALCON provides outstanding performance and a comprehensive support of multi-core platforms, MMX, and SSE2. It serves all industries by a library of more than 1400 operators for blob analysis, morphology, pattern matching, measuring, identification, and 3D vision, to name just a few.

http://www.mvtec.com/halcon/

Aplikasi Perangkat Lunak Vision

COGNEX (http://www.cognex.com/Main.aspx) Vision Systems : All-in-one systems that combine camera, processor and vision software into a

single rugged package, with a simple and flexible user interface for configuring your application.

Vision Software : Vision software gives you the most flexibility for combining the full library of powerful Cognex vision tools with the cameras, frame grabbers and peripherals of your choice, and enables easy integration with PC-based data and control programs.

Vision Sensors : Easy, affordable sensors that can be used in place of photoelectric sensors for more reliable inspection, error-proofing and part detection.

Industrial ID : Fast, reliable 1D and 2D code reading and verification for direct part mark or high-contrast applications.

Industry-Specific Products: A result of over 25 years of vision experience solving the most difficult vision applications, these products include wafer identification, surface mount device placement guidance, cylindrical product inspection and more.

Web and Surface Inspection : Industry-leading technology for detecting and classifying defects during the continuous production of metals, paper, nonwovens, plastics and glass.

Vehicle

wheel

Animal

leg

head Four-legged

Mammal

Move on road

Facing right

Can run, jump

Is herbivorous

Facing right

Penggunaan vision Hari Ini

Contoh state-of-the-art

Damar Darbito, 2013 - Inspeksi Produksi Kartu Seluler

Industrial Vision Penggunaan vision Hari Ini

Industrial Vision Penggunaan vision Hari Ini

Benyamin, 2013 - Inspeksi Produksi Botol Susu plastik

Deteksi kecacatan pada mulut botol

Deteksi kecacatan dalam botol Deteksi kecacatan pinggir botol

61

Recovery 3D layout dan context

BED


Editing images as if they were 3D scenes


Earth viewers (3D modeling)

Image from Microsoft’s Virtual Earth

(see also: Google Earth)


http://www.microsoft.com/virtualearth/

http://earth.google.com/

65

Building Rome in a Day: Agarwal et al. 2009

3D from thousands of images

Hoiem Efros Hebert SIGGRAPH 2005

3D from one image


66

Digit recognition, AT&T labs

http://www.research.att.com/~yann/

Technology to convert scanned docs to text

• If you have a scanner, it probably came with OCR software

License plate readers http://en.wikipedia.org/wiki/Automatic_number_plate_recognition

Optical character recognition (OCR) Penggunaan vision Hari Ini

http://www.research.att.com/~yann

http://en.wikipedia.org/wiki/Automatic_number_plate_recognition

Many new digital cameras now detect faces

Canon, Sony, Nikon …

Face detection


Sony Cyber-shot® T70 Digital Still Camera

Smile detection? Penggunaan vision Hari Ini

http://www.sonystyle.com/webapp/wcs/stores/servlet/ProductDisplay?catalogId=10551&storeId=10151&productId=8198552921665200469&langId=-1



69

Object recognition (in supermarkets)

LaneHawk by EvolutionRobotics

“A smart camera is flush-mounted in the checkout

lane, continuously watching for items. When an

item is detected and recognized, the cashier

verifies the quantity of items that were found under

the basket, and continues to close the transaction.

The item can remain under the basket, and with

LaneHawk, you are assured to get paid for it… “


http://www.evolution.com/products/lanehawk/

70

“How the Afghan Girl was Identified by Her Iris Patterns” Read the story

wikipedia

Vision-based biometrics Penggunaan vision Hari Ini

http://www.cl.cam.ac.uk/~jgd1000/afghan.html

http://en.wikipedia.org/wiki/Afghan_Girl_%28photo%29

71

Fingerprint scanners on

many new laptops,

other devices

Face recognition systems now

beginning to appear more widely http://www.sensiblevision.com/

Login without a password… Penggunaan vision Hari Ini

http://www.sensiblevision.com/

Point & Find, Nokia

Google Goggles

Object recognition (in mobile phones)


http://www.infoworld.com/article/07/04/24/HNnokiasiliconvalley_1.html

http://research.nokia.com/researchteams/vcui/index.html

http://www.google.com/mobile/goggles/#text

The Matrix movies, ESC Entertainment, XYZRGB, NRC

Special effects: shape capture Penggunaan vision Hari Ini

Pirates of the Carribean, Industrial Light and Magic

Special effects: motion capture Penggunaan vision Hari Ini

Based-on Ega Hegarini 2015 - Motion Analysis for sport science

Special effects: motion capture Penggunaan vision Hari Ini

76

Sports

Sportvision first down line

Nice explanation on www.howstuffworks.com

http://www.sportvision.com/video.html


http://www.howstuffworks.com/first-down-line.htm

http://www.howstuffworks.com/

http://www.sportvision.com/video.html

Mobileye

Vision systems currently in high-end BMW, GM, Volvo models

By 2010: 70% of car manufacturers.

Slide content courtesy of Amnon Shashua

Smart cars Penggunaan vision Hari Ini

http://www.mobileye.com/

Smart Vision Drone Penggunaan vision Hari Ini

http://www.nytimes.com/2010/10/10/science/10google.html?ref=artificialintelligence

Google cars Penggunaan vision Hari Ini

http://www.nytimes.com/2010/10/10/science/10google.html?ref=artificialintelligence

Object Recognition: http://www.youtube.com/watch?feature=iv&v=fQ59dXOo63o

Mario: http://www.youtube.com/watch?v=8CTJL5lUjHg

3D: http://www.youtube.com/watch?v=7QrnwoO1-8A

Robot: http://www.youtube.com/watch?v=w8BmgtMKFbY

Interactive Games: Kinect Penggunaan vision Hari Ini

http://www.youtube.com/watch?feature=iv&v=fQ59dXOo63o

http://www.youtube.com/watch?v=8CTJL5lUjHg

http://www.youtube.com/watch?v=8CTJL5lUjHg

http://www.youtube.com/watch?v=7QrnwoO1-8A




http://www.youtube.com/watch?v=w8BmgtMKFbY

Vision systems (JPL) used for several tasks

• Panorama stitching

• 3D terrain modeling

• Obstacle detection, position tracking

• For more, read “Computer Vision on Mars” by Matthies et al.

NASA'S Mars Exploration Rover Spirit captured this westward view from atop

a low plateau where Spirit spent the closing months of 2007.

Vision in space Penggunaan vision Hari Ini

http://www.ri.cmu.edu/pubs/pub_5719.html

http://marsrovers.jpl.nasa.gov/gallery/images.html

Vision-guided robots position nut runners on wheels

Industrial robots


http://www.robocup.org/

NASA’s Mars Spirit Rover

http://en.wikipedia.org/wiki/Spirit_rover

Saxena et al. 2008

STAIR at Stanford

Mobile robots


UNSW_CMU.mpg

http://www.robocup.org/

http://upload.wikimedia.org/wikipedia/commons/d/d8/NASA_Mars_Rover.jpg

http://en.wikipedia.org/wiki/Spirit_rover

http://stair.stanford.edu/


Image guided surgery

Grimson et al., MIT 3D imaging

MRI, CT

Medical imaging

http://groups.csail.mit.edu/vision/medical-vision/surgery/surgical_navigation.html

86

Entertainment : Video Mapping

www.artisuniversalis.com/educational

1. Uses projection to place videographics on a

physical object.

2. Creates an optical illusion using light.

3. Transforms ordinary objects into magical living

entities.


Hari ini sudah sama-sama kita bicarakan :

Definisi

Dasar Ilmu yang harus dikuasai

Tantangannya

Anatominya

Implementasi Computer Vision dalam kehidupan

Selanjutnya ?

Terserah anda… (mau jadi player?

Atau mau jadi penonton saja?)

Kesimpulan

Terima Kasih

Thank You

Question?

merci δ

Itot

Computer Vision - Gunadarma...

Documents

Transcript of Computer Vision - Gunadarma...