FP7 Project Grant Number 288238 Brief introduction of WP4 10min - - PowerPoint PPT Presentation

▶
fp7 project grant number 288238 brief introduction of wp4
SMART_READER_LITE
LIVE PREVIEW

FP7 Project Grant Number 288238 Brief introduction of WP4 10min - - PowerPoint PPT Presentation

Mar 29, 2023 444 likes •1.17k views

FP7 Project Grant Number 288238 Brief introduction of WP4 10min General objectives and role in VENTURI Partners, tasks, deliverables, milestones Timeline Main results (first year) 25min Plans for second year 15min Conclusions

slide-1
SLIDE 1

FP7 Project Grant Number 288238

slide-2
SLIDE 2

» Brief introduction of WP4

˃ General objectives and role in VENTURI ˃ Partners, tasks, deliverables, milestones ˃ Timeline

» Main results (first year) » Plans for second year » Conclusions and questions

2 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

10min 25min 15min

slide-3
SLIDE 3

3 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

WP5: Adaptive Content Harvesting, Creation and Delivery WP4: Context Sensing and Interpretation WP3: User Interface and Interaction Design WP1: Project Management WP6 Technology Integration, Evaluation and Test-Cases WP7: Dissemination and Exploitation WP2: Architecture Definition and Development

slide-4
SLIDE 4

4 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-5
SLIDE 5

5 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

On-board sensors:

  • vision: camera(s)
  • audio: microphone(s)
  • location: GPS, …
  • relative orientation / motion:

accelerometers, gyroscopes

  • absolute orientation: compass
  • ther: pressure, proximity

Data sources:

  • maps
  • DEMs
  • geo-tagged data: images,

3D models, etc.

slide-6
SLIDE 6

6 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-7
SLIDE 7

7 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-8
SLIDE 8

8 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-9
SLIDE 9

9 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-10
SLIDE 10

10 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Absolute orientation (yaw, pitch, roll, field of view, etc.) Geographical localization (lat/lon/alt) (also indoor)

slide-11
SLIDE 11

11 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

  • Scene reconstruction
  • Model the immediate

surroundings of the mobile device

slide-12
SLIDE 12

12 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

  • User behavior understanding
  • What’s happening here?
  • Noise level / scene classification
  • Infer what the user is doing and

what he could experience with his senses

slide-13
SLIDE 13

13 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Enable and explore innovative ways to interact with the environment

Example: SixthSense (Fluid Interface Group at MIT Media Lab)

slide-14
SLIDE 14

14 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

FBK 28MM ST-Italy 21MM ST-France Fraunhofer HHI 19MM metaio 16MM INRIA 18MM Sony 6MM STE 3MM

Total WP4 MMs: 111

slide-15
SLIDE 15

15 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

T4.1 Hardware sensor interpretation T4.2 Auditory scene analysis T4.3 Visual analysis of the environment and geo-position refining T4.4 Physical surroundings modeling T4.5 What is, or could be, going on around me?

slide-16
SLIDE 16

16 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4 M1 = oct 2011 M12 = sep 2012 M24 = sep 2013 M36 = sep 2014

T4.1 T4.2 T4.3 T4.4 T4.5

M18 = mar 2013 M6 = mar 2012 M15 = dec 2012 M30 = mar 2014 M9 = jun 2012

first year

slide-17
SLIDE 17

17 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» D4.1 Synchronized dataset containing a dump of all on-board

sensors of a real device, while being used in a real environment (STE, Mar 2012, M6, O-CO)

» D4.2 Report on expected platform requirements of WP4

algorithms (FBK, Jun 2012, M9, R-CO)

» D4.3 WP4 outcome definitions and API specification for inter-

task / inter-WP communications (ST-Italy, Dec 2012, M15, R-CO)

» D4.4 Prototype of and Report on multi-modal estimation of geo-

location and orientation (FBK, Mar 2013, M18, P-CO / R-PU)

» D4.5 In-the-lab prototype and Report on visual analysis and

reconstruction of the scene (Fraunhofer, Sep 2013, M24, P-CO / R- PU)

» D4.6 Final prototype integrating all WP4 outcomes (FBK, Mar

2014, M30, P-CO)

first year

slide-18
SLIDE 18

» MS9 First version of chain for geo-localization / orientation estimation, based

  • n non-audio / non-video (FBK, Sep 2012, M12)

» MS10 Final version of chain for geo-localization / orientation estimation

(FBK, Mar 2013, M18)

» MS11 WPC version of visual analysis and reconstruction of the scene, ready

for integration (FBK, Sep 2013, M24)

18 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

first year

slide-19
SLIDE 19

19 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4 M1 = oct 2011 M12 = sep 2012 M24 = sep 2013 M36 = sep 2014

T4.1 T4.2 T4.3 T4.4 T4.5

M18 = mar 2013 M6 = mar 2012 M15 = dec 2012 M30 = mar 2014 M9 = jun 2012

D4.1 D4.2 D4.3 D4.4 D4.5 D4.6

first year

MS9 MS10 MS11

slide-20
SLIDE 20

» Brief introduction of WP4

˃ General objectives and role in VENTURI ˃ Partners, tasks, deliverables, milestones ˃ Timeline

» Main results (first year) » Plans for second year » Conclusions and questions

20 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

10min 25min 15min

slide-21
SLIDE 21

21 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-22
SLIDE 22

» Cross-partner / Cross-task results

˃ Shared tools and Android OS improvements ˃ Deliverables

» Results showcased in Y1 demo

˃ Algorithms and their integration

» Other results (relevant for Y2 and Y3 Use Cases)

˃ Analysis of state of the art ˃ Design, development, implementation of algorithms ˃ Evaluation and testing

22 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-23
SLIDE 23

23 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» One of the main requirement elicited (in T2.1.2): accurate and common timestamping between camera and sensor domain

ST9500-based platform: done ST9540-based platform: done, under debugging Next (8680) platform: taken into account from the beginning

slide-24
SLIDE 24

24 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Video: timestamping

slide-25
SLIDE 25

Android porting layer Android libraries

Kernel space User space

Android application Access services Android application framework Core Multimedia Connectivity Android runtime

Android

  • rigin

ST-Ericsson proprietary Non-Android FOSS

Legend:

Sensor HAL Sensor Service (server) Sensor Manager library (client) Sensor Manager Class Sensor Manager JNI I2C Core I2C driver Application Sensor data are Time stamped here

Third- party/other Ambient Light driver Proximity driver Gyro-scope driver Magneto- meter driver Accelero- meter driver

Sensors

I2C

Hardware

Sensor HAL provides Timestamp here Sensor drivers (Linux) not providing Timestamps Imaging Signal Processor

Android ICS Vanilla

slide-26
SLIDE 26

Android porting layer Android libraries

Kernel space User space

Android application Access services Android application framework Core Multimedia Connectivity Android runtime

Android

  • rigin

ST-Ericsson proprietary Non-Android FOSS

Legend:

Sensor HAL Sensor Service (server) Sensor Manager library (client) Sensor Manager Class Sensor Manager JNI I2C Core I2C driver Application Sensor data are Time stamped here

Third- party/other Ambient Light driver Proximity driver Gyro-scope driver Magneto- meter driver Accelero- meter driver

Sensors

I2C

Hardware

Sensor HAL provides Timestamp here Time comes from here SIA IIO Framework

VENTURI modification for 9500/9540 platforms

slide-27
SLIDE 27

» Ongoing experiments to quantitatively evaluate the impact of accurate timestamping on actual WP4 algorithms.

27 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-28
SLIDE 28

» FBKSequenceRecorder tool

» It can capture and dump to disk timestamped data from all on-board sensors (raw video, audio, other sensors). » On STE boards, it exploits the new timestamping APIs (still debugging on STE9540). However, it works on any Android device. » Source code and APK shared with VENTURI partners. » Possibility to share it with general public.

28 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-29
SLIDE 29

29 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» D4.1 Synchronized dataset containing a dump of all on-board

sensors of a real device, while being used in a real environment (STE, Mar 2012, M6, O-CO)

» D4.2 Report on expected platform requirements of WP4

algorithms (FBK, Jun 2012, M9, R-CO)

» D4.3 WP4 outcome definitions and API specification for inter-

task / inter-WP communications (ST-Italy, Dec 2012, M15, R-CO)

» D4.4 Prototype of and Report on multi-modal estimation of geo-

location and orientation (FBK, Mar 2013, M18, P-CO / R-PU)

» D4.5 In-the-lab prototype and Report on visual analysis and

reconstruction of the scene (Fraunhofer, Sep 2013, M24, P-CO / R- PU)

» D4.6 Final prototype integrating all WP4 outcomes (FBK, Mar

2014, M30, P-CO)

D4.1

slide-30
SLIDE 30

» Use Case datasets and algorithm specific

  • datasets. GT provided when possible /

relevant. » Immediately shared with VENTURI partners through mailing list. » Integral version on restricted section of VENTURI website:

https://venturi.fbk.eu/reserved/datasets

» General interest datasets shared with the public: https://venturi.fbk.eu/results/public-

datasets

» Submitted document D4.1 needs to be updated with current status. » Website and D4.1 document will periodically be updated with new data during the entire life of VENTURI.

30 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Website

slide-31
SLIDE 31

31 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» D4.1 Synchronized dataset containing a dump of all on-board

sensors of a real device, while being used in a real environment (STE, Mar 2012, M6, O-CO)

» D4.2 Report on expected platform requirements of WP4

algorithms (FBK, Jun 2012, M9, R-CO)

» D4.3 WP4 outcome definitions and API specification for inter-

task / inter-WP communications (ST-Italy, Dec 2012, M15, R-CO)

» D4.4 Prototype of and Report on multi-modal estimation of geo-

location and orientation (FBK, Mar 2013, M18, P-CO / R-PU)

» D4.5 In-the-lab prototype and Report on visual analysis and

reconstruction of the scene (Fraunhofer, Sep 2013, M24, P-CO / R- PU)

» D4.6 Final prototype integrating all WP4 outcomes (FBK, Mar

2014, M30, P-CO)

D4.2

slide-32
SLIDE 32

» Report on expected platform requirements of WP4 algorithms (FBK, Jun 2012, M9, R-CO) » Incremental wrt existing deliverables:

˃ D2.1.1 Use cases, application definition and system requirements for STE U8500-based platform ˃ D2.2.1 Early Detailed Design Specifications for STE U8500-based platform

» Submitted on July, 11th, 2012

˃ Initial identification of algorithmic modules (input/output, impact on requirements). ˃ List of requirements updated wrt D2.1.1 and D2.2.1

32 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-33
SLIDE 33

» Cross-partner / Cross-task results

˃ Shared tools and Android OS improvements ˃ Deliverables

» Results showcased in Y1 demo

˃ Algorithms and their integration

» Other results (relevant for Y2 and Y3 Use Cases)

˃ Analysis of state of the art ˃ Design, development, implementation of algorithms ˃ Evaluation and testing

33 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-34
SLIDE 34

First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-35
SLIDE 35

35 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-36
SLIDE 36

36 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-37
SLIDE 37

» Map reconstruction -> .3dmap file » Indoor and small-scale environments (table-top scenarios) with roughly one meter cube volume. » Development and testing also for larger scale environments for UCY2 and UCY3.

37 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-38
SLIDE 38

» .3dmap file -> 3D markerless tracking

38 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-39
SLIDE 39

39 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-40
SLIDE 40

» Cross-partner / Cross-task results

˃ Shared tools and Android OS improvements ˃ Deliverables

» Results showcased in Y1 demo

˃ Algorithms and their integration

» Other results (relevant for Y2 and Y3 Use Cases)

˃ Analysis of state of the art ˃ Design, development, implementation of algorithms ˃ Evaluation and testing

40 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-41
SLIDE 41

41 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4 M1 = oct 2011 M12 = sep 2012 M24 = sep 2013 M36 = sep 2014

T4.1 T4.2 T4.3 T4.4 T4.5

M18 = mar 2013 M6 = mar 2012 M15 = dec 2012 M30 = mar 2014 M9 = jun 2012

D4.1 D4.2 D4.3 D4.4 D4.5 D4.6

first year

MS9 MS10 MS11

slide-42
SLIDE 42

42 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Audio Patterns Database FFT Transform Ambiant audio Classifier OSM Data Analysis Results Feature Extractor

Filter possible patterns using OSM context Office Restaurant Parking … other places Car Phone ringing … other objects

slide-43
SLIDE 43

43 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Name Type (scene or object)

OSM Tags OSM Tags

OSM Tags Feature Vector Feature Vector Feature Vector

…

Features

Feature vectors based

  • n

mean frequency power repartition over time

Fingerprints

Classifier uses objects and scene fingerprints stored in XML format

slide-44
SLIDE 44

44 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Example: Airbus A319

An Airbus A319 when seating near the aisle

Example: restaurant

Busy restaurant

slide-45
SLIDE 45

» Source code released to partners » Cross-platform C++ API shared with partners » Ready for integration with Android NDK » Fully documented (also in D4.3) » Can perform real-time and offline classification

  • f environment areas and objects

45 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-46
SLIDE 46

» implementation of a pedometer based on walking models and vertical acceleration patterns » estimation of orientation based on gyro data and an OpenStreetMap network » estimation of position based on estimated stride length, orientation and OpenStreetMap network

46 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-47
SLIDE 47

47 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Input

HEADING DRIFT COMPENSATION

Pedestrian Dead-Reckoning

Navigation network Output

PEDOMETER

Position Heading

GYROMETRY DISTANCE COMPENSATION

heading position

INTERNAL VARIABLES Map-Matching

slide-48
SLIDE 48

48 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Distance errors Map Matching needed Heading errors due to drift HDC needed Distance and Heading errors Map Matching + HDC needed

slide-49
SLIDE 49

49 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Video: PDR

slide-50
SLIDE 50

» Scene classification (also known as Visual

categorization) aims to automatically classify images into a set of different semantic classes, based on information extracted from the images. » holistical technique, based on features extracted from the YUV images (global features, SVM).

50 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Target classes: {Building, Open Country, Forest} Building

Image representation Classification model

slide-51
SLIDE 51

51 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» Downloaded a publicly available dataset, in order to have a benchmark vs state of the art technique (“Modeling the shape of the scene: a holistic representation of the spatial envelope” - Aude Oliva, Antonio Torralba); » 2668 images in 8 classes;

coast forest highway inside city mountain

  • pen country

street tall building

slide-52
SLIDE 52

52 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Accuracy: 83.3%; Torralba-Oliva obtain 83.7% of accuracy with a much more complex approach (using FFT);

slide-53
SLIDE 53

First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Video: SceneClass

slide-54
SLIDE 54

54 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-55
SLIDE 55

55 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Video: alignFBK

slide-56
SLIDE 56

56 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4 Camera image 3D model

?

Goal: register known 3D model with 2D camera image

slide-57
SLIDE 57

57 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» 2D-3D matching for pose estimation

˃ Exploitation of architectural properties of man-made buildings, e.g. by extraction of vanishing lines / points ˃ Use of inertial sensor data and camera intrinsic calibration to constrain the calculation of vanishing points

[Buhrig, Eisert: “Pose estimation for urban environments”, CVMP, Dec. 2012]

slide-58
SLIDE 58

58 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» 2D-3D matching for pose estimation

˃ Example: register 3D model of tax office in image ˃ Result: exact position and orientation of 3D model

Estimated orientation with vanishing cues Position and

  • rientation refinement
slide-59
SLIDE 59

59 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Camera pose estimation Map creation and update Features detector Temporal tracking Video Analytics Camera position and Sparse 3d map GEOMETRY PIPELINE

Structure from Motion

Goal: estimating the trajectory of a moving camera and reconstruct the shape of the surrounding environment

slide-60
SLIDE 60

60 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» Android app developed to easy handle camera properties and processing. » Incremental algorithm porting and testing using JNI API to integrate c/cpp code in Android. » Developed an unique build system solution to target X86/ARM/Android platform. » Tested application in indoor/outdoor.

Features Detection: indoor

  • utdoor

Optical flow indoor

Demo

slide-61
SLIDE 61

61 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

» User-generated 3D content

˃ Take “3D snapshots” of the world with mobile camera ˃ Dense reconstruction from high-resolution stereo-images, allows recovering object surfaces at high detail ˃ Example: reconstruction from 2 freehand-shots in a museum (uncalibrated setup, uncontrolled illumination)

slide-62
SLIDE 62

62 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Indoor example

slide-63
SLIDE 63

63 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

Outdoor example:

reconstruction from 2 free-hand shots (uncalibrated setup, uncontrolled illumination)

slide-64
SLIDE 64

» Text-in-scene detection and recognition (T4.3.3). » Robust detection of interest points / regions (T4.3.1). » Real-time terrain model rendering on a mobile device (T4.3.2). » Gesture recognition (T4.4). » Generation of a sparse 3D map, single camera (T4.4). » Image match with google StreetView

64 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-65
SLIDE 65

» Cross-partner / Cross-task results

˃ Timestamping (T4.1) ˃ Acquisition tool ˃ Deliverables D4.1, D4.2

» Results showcased in Y1 demo

˃ Map reconstruction (T4.4) ˃ 3D markerless tracking (T4.3)

» Other results (Y2 and Y3 Use Cases)

˃ Auditory scene analysis (T4.2) ˃ PDR-based localization (T4.1) ˃ Scene Classification (T4.3.1) ˃ Terrain model alignment (T4.3.2) ˃ 2D-3D model building matching (T4.3.2) ˃ Sparse 3D map – SfM (T4.4) ˃ Dense 3D reconstruction (T4.4) ˃ SoTA analysis, other preliminary works

65 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-66
SLIDE 66

» Brief introduction of WP4

˃ General objectives and role in VENTURI ˃ Partners, tasks, deliverables, milestones ˃ Timeline

» Main results (first year) » Plans for second year » Conclusions and questions

66 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

10min 25min 15min

slide-67
SLIDE 67

67 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

T4.1 T4.2 T4.3 T4.4 T4.5 progress

slide-68
SLIDE 68

68 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

T4.1 T4.2 T4.3 T4.4 T4.5

M12 M12 M12 M12 M12

progress

slide-69
SLIDE 69

» Complete D4.3: WP4 outcome definitions and API specification for inter-task/inter-WP communications (ST-I, M15 Dec2012, R-CO)

˃ Define all algorithmic modules: brief description, input/output, APIs ˃ Following expected UCY2 and UCY3, identify interactions, bottlenecks, role of T4.5 (integration)

» Complete T4.1, T4.3, T4.4, following specifications defined in D4.3, leading to D4.5. » Implement UCY2, integrating and fusing appropriate WP4 outcomes (T4.5)

69 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4

slide-70
SLIDE 70

70 First Review Meeting. Hosted by Fraunhofer HHI in Berlin, Germany on the 18th of December 2012 WP4 M1 = oct 2011 M12 = sep 2012 M24 = sep 2013 M36 = sep 2014

T4.1 T4.2 T4.3 T4.4 T4.5

M18 = mar 2013 M6 = mar 2012 M15 = dec 2012 M30 = mar 2014 M9 = jun 2012

D4.1 D4.2 D4.3 D4.4 D4.5 D4.6

first year

MS9 MS10 MS11