Wrap Up Lecture Instructor - Simon Lucey 16-423 - Designing - - PowerPoint PPT Presentation

▶
wrap up lecture
SMART_READER_LITE
LIVE PREVIEW

Wrap Up Lecture Instructor - Simon Lucey 16-423 - Designing - - PowerPoint PPT Presentation

Sep 14, 2023 165 likes •778 views

Wrap Up Lecture Instructor - Simon Lucey 16-423 - Designing Computer Vision Apps Today Review - Project Presentations Emerging Trends in Mobile Vision Facial Feature Detection in iOS. Project Discussions Project Presentations

slide-1
SLIDE 1

Wrap Up Lecture

Instructor - Simon Lucey

16-423 - Designing Computer Vision Apps

slide-2
SLIDE 2

Today

  • Review - Project Presentations
  • Emerging Trends in Mobile Vision
  • Facial Feature Detection in iOS.
  • Project Discussions
slide-3
SLIDE 3

Project Presentations

  • Each team will have 2.5 minutes per member to present

(e.g. 2 member team will have 5 minutes allotted).

  • We will be strict with this, and will start class at exactly

12pm, please be early!!!

  • Each team will upload their YouTube clips to the link.
slide-4
SLIDE 4

Project Presentations.

  • 16423 staff will select the “best” presentation plus an

honorable mention.

  • The result will be announced on Piazza after class, the

winning team can come by my office to pick up their prize.

  • All details on the Final Project submission can be found at:-

http://16423.courses.cs.cmu.edu/slides/Final_Project.pdf .

  • Final Write Up is due on midnight Friday 11th of December.

Reminder: Final Project makes up 50% of your grade!!!

slide-5
SLIDE 5

Today

  • Review - Project Presentations
  • Emerging Trends in Mobile Vision
  • Facial Feature Detection in iOS.
  • Project Discussions
slide-6
SLIDE 6

Why is Mobile CV Different?

6

slide-7
SLIDE 7

Why is Mobile CV Different?

6

slide-8
SLIDE 8

Why is Mobile CV Different?

7

slide-9
SLIDE 9

Why is Mobile CV Different?

7

slide-10
SLIDE 10

Why is Mobile CV Different?

7

slide-11
SLIDE 11

Balancing Power versus Perception

8

slide-12
SLIDE 12

Emerging Trends - Low Power

9

Taken from: http://lpirc.net/

slide-13
SLIDE 13

Emerging Trends - High Speed Camera

10

iPhone 6 Samsung Galaxy S5

slide-14
SLIDE 14

Emerging Trends - High Speed Camera

10

iPhone 6 Samsung Galaxy S5

slide-15
SLIDE 15

Emerging Trends - Depth Cameras

11

slide-16
SLIDE 16

Emerging Trends - Augmented Reality

12

slide-17
SLIDE 17

Emerging Trends - Mobile SLAM

13

  • P. Tanskanen, K. Kolev, L. Meier, F. Camposeco, O. Saurer, M. Pollefeys : Live metric 3d reconstruction on mobile phones. (ICCV 2013)
slide-18
SLIDE 18

Emerging Trends - Mobile SLAM

13

  • P. Tanskanen, K. Kolev, L. Meier, F. Camposeco, O. Saurer, M. Pollefeys : Live metric 3d reconstruction on mobile phones. (ICCV 2013)
slide-19
SLIDE 19

Emerging Trends - Mobile SLAM

13

  • P. Tanskanen, K. Kolev, L. Meier, F. Camposeco, O. Saurer, M. Pollefeys : Live metric 3d reconstruction on mobile phones. (ICCV 2013)
slide-20
SLIDE 20

New Work….

14

slide-21
SLIDE 21

New Work….

14

slide-22
SLIDE 22

Emerging Trends - Deep Vision

15

slide-23
SLIDE 23

Emerging Trends - Deep Learning

ImageNet Challenge Year

BC

(before ConvNets)

AD

(after deep learning)

6.8%

slide-24
SLIDE 24

Emerging Trends - Deep Learning

  • Data and Compute Speed Matter
  • Energy and Space do NOT!!!

Training

slide-25
SLIDE 25

Emerging Trends - Deep Learning

  • Data and Compute Speed Matter
  • Energy and Space now Matter!!!!

Testing

slide-26
SLIDE 26

Emerging Trends - Deep Learning

  • Data and Compute Speed Matter
  • Energy and Space now Matter!!!!

Testing

slide-27
SLIDE 27

Reminder: ASICs for Low Energy

  • Application Specific Integrated Circuits (ASIC)
  • ASICs are perfect for targeting a specific application domain.
  • Inherently low-power as they are “frozen in silicon” for a

specific application domain (e.g. graphics cards, ethernet cards, DSPs, etc.).

  • Drawbacks,
  • incredibly expensive to develop.
  • time consuming and resource-intensive to develop.
  • Positives,
  • Extremely energy efficient.

19

slide-28
SLIDE 28

Emerging Trends - “Frozen” DeepNets

20

(Taken from recent talk by Yann Lecunn at Hot Chips conference in May 2015).

slide-29
SLIDE 29

Emerging Trends - Deep Learning

21

APIs in the current versions of OpenGL ES do not have the “scatter”

slide-30
SLIDE 30

Emerging Trends - Deep Learning

21

APIs in the current versions of OpenGL ES do not have the “scatter”

Deep Learning

Check out this article.

slide-31
SLIDE 31

Emerging Trends - Deep Learning

22

(Taken from recent talk by Yann Lecunn at Hot Chips conference in May 2015).

slide-32
SLIDE 32

Today

  • Review - Project Presentations
  • Emerging Trends in Mobile Vision
  • Facial Feature Detection in iOS.
  • Project Discussions
slide-33
SLIDE 33

Face Detection

slide-34
SLIDE 34

Face Detection in iOS

  • In iOS face detection comes built in, and can be performed

much more efficiently than standard OpenCV.

  • Utilizes the QuartzCore and CoreImage frameworks within

the project.

slide-35
SLIDE 35

Face Detection in iOS

  • In iOS face detection comes built in, and can be performed

much more efficiently than standard OpenCV.

  • Utilizes the QuartzCore and CoreImage frameworks within

the project.

slide-36
SLIDE 36

Facial Feature Detection

slide-37
SLIDE 37

CIFaceFeature

slide-38
SLIDE 38

CIFaceFeature

slide-39
SLIDE 39

CIFaceFeature Example

  • We are now going to demonstrate a simple example of face

detection in iOS.

  • On your browser please go to the address,

https://github.com/slucey-cs-cmu-edu/CIFaceFeature_Lena

  • Or better yet, if you have git installed you can type from the

command line. $ git clone https://github.com/slucey-cs-cmu-edu/CIFaceFeature_Lena.git

slide-40
SLIDE 40

CIFaceFeature Example

slide-41
SLIDE 41

CIFaceFeature Example

slide-42
SLIDE 42

CIFaceFeature Example

slide-43
SLIDE 43

CIFaceFeature Example

slide-44
SLIDE 44

CIFaceFeature Example

slide-45
SLIDE 45

CIFaceFeature Example

slide-46
SLIDE 46

Smerk and GPUImage

  • Recently, an extension to GPUImage was proposed to allow

for the utilization of iOS face detection within iOS.

  • Called “Smerk” - GitHub project page can be found at:-

https://github.com/MattFoley/Smerk

slide-47
SLIDE 47

Smerk and GPUImage

  • Recently, an extension to GPUImage was proposed to allow

for the utilization of iOS face detection within iOS.

  • Called “Smerk” - GitHub project page can be found at:-

https://github.com/MattFoley/Smerk

slide-48
SLIDE 48

Smerk Example

  • We are now going to demonstrate how we can perform real-

time face tracking through GPUImage.

  • On your browser please go to the address,

https://github.com/slucey-cs-cmu-edu/Smerk_Example

  • Or better yet, if you have git installed you can type from the

command line. $ git clone https://github.com/slucey-cs-cmu-edu/Smerk_Example.git

slide-49
SLIDE 49

Smerk Example

slide-50
SLIDE 50

Smerk Example

slide-51
SLIDE 51

Smerk Example

slide-52
SLIDE 52

Smerk Example

slide-53
SLIDE 53

Smerk Example

slide-54
SLIDE 54

Smerk Example

slide-55
SLIDE 55

Smerk Example

slide-56
SLIDE 56

Smerk Example

slide-57
SLIDE 57

Smerk Example

slide-58
SLIDE 58

Smerk Example

slide-59
SLIDE 59

Today

  • Review - Project Presentations
  • Emerging Trends in Mobile Vision
  • Facial Feature Detection in iOS.
  • Project Discussions
Core Hubs:
All Converters PDF Hub Video Hub Audio Hub Image Hub File Compressor