Application on MobilityFirst Winlab Summer Internship 2015 - - PowerPoint PPT Presentation

▶
application on mobilityfirst
SMART_READER_LITE
LIVE PREVIEW

Application on MobilityFirst Winlab Summer Internship 2015 - - PowerPoint PPT Presentation

Feb 12, 2023 256 likes •409 views

Real-Time Cyber Physical Systems Application on MobilityFirst Winlab Summer Internship 2015 Karthikeyan Ganesan, Wuyang Zhang, Zihong Zheng Shantanu Ghosh, Avi Cooper WINLAB SUMMER 2015 TEAM M MEMBER BERS Karthi hikey eyan an Ganes

slide-1
SLIDE 1

Real-Time Cyber Physical Systems Application on MobilityFirst

Winlab Summer Internship 2015 Karthikeyan Ganesan, Wuyang Zhang, Zihong Zheng Shantanu Ghosh, Avi Cooper

slide-2
SLIDE 2

TEAM M MEMBER BERS

WINLAB SUMMER 2015

Karthi hikey eyan an Ganes esan an Wuyang ang Zhang ng Zihong

  • ng Zheng

ng Shant antanu anu Ghosh Avi Cooper er

slide-3
SLIDE 3

PRE RELIM IMINARY INARY GOAL AL OF OUR UR PRO ROJECT ECT

WINLAB SUMMER 2015

Mobil ilityFi ityFirst rst Virtua tual Ne Networ work

Client ent side: de:

Run an insta tance nce of camera ra syste stem; m; Transmi nsmits ts video

  • in

standard rd forma mat; t; Simple le graphical hical interf rface ce to display ay results lts

Serv rver r side: e:

Imple lement ment server ver applicatio cation for object ct recogn gniti ition

  • n;

Return rn the result lt

CPS Application based on MF

slide-4
SLIDE 4

CU CURR RRENT NT FRA RAME

WINLAB SUMMER 2015

Application: Tried to debug some MF Android applications done before(mfstack, mfping, etc.) Learning how to set up the WIFI access point on Router to enable the MF connection. Image Recognition: Done essential time analysis for different phases in Image Matching. Have some progress on Speed optimization. Also working on strategies to improve Accuracy. Cloud Computing: Set up the Hadoop Cluster, constructed by Master and Slaves that could run Map & Reduce jobs. Now working on Hadoop Image Processing Interface(HIPI).

slide-5
SLIDE 5

IMAG AGE PRO ROCES CESSO SOR

WINLAB SUMMER 2015

  • Server Startup
  • Load Descriptors into memory
  • Image Matching
  • Isolate Descriptors from test image
  • Find Matching Descriptors from Database using the Flann Class
  • Algorithm to determine Object from matched descriptors
slide-6
SLIDE 6

IMAG AGE PRO ROCES CESSO SOR R OPTIM TIMIZ IZATI ATION: N: SPEED

WINLAB SUMMER 2015

  • Server Startup
  • Load Descriptors into memory
  • Image Matching
  • Isolate Descriptors from test image
  • Find Matching Descriptors from Database
  • Build a KD tree
  • Perform a knn search
  • Algorithm to determine Object from matched descriptors
slide-7
SLIDE 7

IMAG AGE PRO ROCES CESSO SOR R OPTIM TIMIZ IZATI ATION: N: SPEED

WINLAB SUMMER 2015

500 1000 1500 2000 2500 3000 3500 4000 4500 50 100 250 500 1000

Search h Time (ms ms) Image Databas ase e Size

Old

5 10 15 20 25 30 35 40 50 100 250 500 1000

Search h Time (ms ms) Image Databas ase e Size

New

slide-8
SLIDE 8

IMAG AGE PRO ROCES CESSO SOR R OPTIM TIMIZ IZATI ATION: N: SPEED

WINLAB SUMMER 2015

  • Descriptor Isolation
  • Largest portion of computing time.
  • Fixes :
  • Try utilizing the GPU to perform SURF
  • Try a method that better utilizes multiple CPU cores
  • Try different feature detector other than SURF
slide-9
SLIDE 9

IMAG AGE PRO ROCES CESSO SOR R OPTIM TIMIZ IZATI ATION: N: ACC CCUR URACY CY

WINLAB SUMMER 2015

  • True Positives
  • Detects objects in test image that exist in Database with close to 100%

accuracy

  • False Positives
  • Often Falsely recognize objects that don’t exist in the image (Close to

50%)

  • Fix
  • Improve the Algorithm used to determine object from Matched

Descriptors with additional checks to confirm the actual existence of the

  • bject.
slide-10
SLIDE 10

ANDROID PHONE

WINLAB SUMMER 2015

MFStack is used to install the MF stack on phone. Also a launcher to start and stop the mf stack service. Devices under MF network is actually communicating through the MF stack. MFPing achieves the basic ping function for MF such as the ping for TCP/IP. We plan use it to test the MF connection after we set up the access point

  • n router.
slide-11
SLIDE 11

PLAN TO SET UP THE CONNECTION

WINLAB SUMMER 2015

Master node as Server manages/allocate image recognition jobs Glass captures video Android phone handles networking

Bluetooth MF

Access Point on router

MF HADOOP

Slaves nodes

ORBIT outdoor nodes

slide-12
SLIDE 12

WINLAB SUMMER 2015

Background: Optimized image processing algorithm, current system process 1,000 images within 113 ms. Impediment: Load the large size of descriptors in the database into a local memory before matching images. 1,000 images : 1 GB, linear accumulation -> 10,000 images : 10 GB HIPI: Image processing interface under the framework of HADOOP. Distribute the database into several machines so as to support over 10,000 images. HIPI FRAMEWORK

Hadoop Image Processing Interface

slide-13
SLIDE 13

Ne Next t Week k Plan

WINLAB SUMMER 2015

Set et up the he conn nnection ection bet etwee een n And ndroid

  • id Pho

hone ne an and orbi bit t nodes through gh an access ss point. t. Conti tinue develop p the client progra gram based d on google gle glass ss and Androi

  • id

d phone. Use e HI HIPI I to em embed edde ded d Image age Proc

  • cess

essin ing Prog

  • gra

ram m an and Ha Hadoo

  • op

p Frame mework. work.

slide-14
SLIDE 14

Questions?