Video Processing and Understanding Lab
Universidad Autónoma de Madrid, Escuela Politécnica Superior

TEC2014-53176-R HAVideo (2015-2017-2018)
High Availability Video Analysis for People Behaviour Understanding

Supported by the
Ministerio de Economía y Competividad
of the Spanish Goverment
Workpackages

WP1: Video Analysis Framework

This work package aims at the initial establishment and maintenance of a development framework for the remaining work packages.

T.1.1 (M1-36): Update, upgrade and maintenance of the system infrastructure.

Milestones:

  • MS1.1.1: Availability of a portable (wireless) infrastructure for sequence acquisition and computing (M6 and updates every 6 months if required).

T.1.2 (M1-36): Update and maintenance of the Distributed Video Analysis (DiVA) software architecture.

Milestones:

  • MS1.2.1: Update DiVA framework for contextual information (M6)
  • MS1.2.2: Update DiVA framework for collaborative processing (M24 and updates every 6 months if required).

T.1.3 (M1-36): Update and maintenance of a camera network simulatoir.

Milestones:

  • MS1.3.1: Definition of requirements and adaptation of an existing simulator (M2)
  • MS1.3.2: Simulator upgrade (M6)
  • MS1.3.3: Integration of 3D environment (M12).

WP2: Video analysis tools, models and performance indicators

This work package aims at developing the required tools, models and control signals in order to enable the development of adaptive and collaborative approaches for video-based understanding of people behaviour.

T.2.1 (M1-32): Analysis Tools for human behavior understanding.

Milestones:

  • MS2.1.1: Development and integration of new long-term algorithms for single-camera people behaviour understanding (M12)
  • MS2.1.2: Development and integration of new long-term algorithms for multi-camera people behaviour understanding (M32).

T.2.2 (M1-24): Contextual modeling and extraction. [Extended M1-M30]

Milestones:

  • MS2.2.1: Single-camera context and offline acquisition (M12)
  • MS2.2.2: Online maintenance of single-camera context (M18)
  • MS2.2.3: Multi-camera context and offline acquisition (M18) [rescheduled M27]
  • MS2.2.4: Online maintenance of multi-camera context (M24) [rescheduled M30]

T.2.3 (M5-30): Quality Analysis.

Milestones:

  • MS2.3.1: Online evaluation algorithms for object detection & tracking in people behaviour understanding (M18) [rescheduled M27]
  • MS2.3.2: Online algorithms to evaluate view importance in multi-camera settings (M22) [rescheduled M30]
  • MS2.3.3: Online task relevance evaluation algorithms in multi-camera settings (M30).

T.2.4 (M1-36): Exploration and viability studies.

Milestones:

  • depending on the evolution of the task

WP3: Self-configurable approaches for long-term analysis

This work package aims at developing approaches to enable the high availability of video for applications based on people behaviour understanding. It makes use of the results in WP1 and WP2 to provide self-configurable analysis over time and adapt to changing conditions which frequently occur in long-term analysis. This work package consists of two tasks regarding the proposed strategies to enable self-configuration: adaptation and collaboration.

T.3.1 (M12-34): Adaptation approaches [Delayed to M18-36]

Milestones:

  • MS3.1.1: First version of adaptive algorithms based on contextual information (M18) [rescheduled M24]
  • MS3.1.2: Final version of adaptive algorithms based on contextual information (M30) [rescheduled M34]
  • MS3.1.3: First version of adaptive algorithms based on quality information (M22) [rescheduled M27]
  • MS3.1.4: Final version of adaptive algorithms based on quality information (M34).

T.3.2 (M18-34): Collaborative approaches.

Milestones:

  • MS3.2.1: First version of people behaviour understanding based on quality and context (M24) [rescheduled M27]
  • MS3.2.2: Final version of people behaviour understanding based on quality and context (M34)

WP4: Evaluation framework, demonstrators and dissemination

This work package aims at evaluating and integrating the algorithms developed within WP1, WP2 and WP3, in order to conform the global analysis chain to provide solutions to long-term video analysis for people behaviour understanding. This work package also coordinates the project dissemination activities.

T.4.1 (M1-36): Evaluation framework.

Milestones:

  • MS4.1: Evaluation methodology and datasets (M6 and updates each 6 months if required)

T4.2 (M12-36): Use Cases and Demonstrators [Delayed to M25-36]

Milestones:

  • MS4.2.1: Framework and Guidelines for the development of applications (M15) [rescheduled M26]
  • MS4.2.2: Planning of first set of applications (M15) [cancelled]
  • MS4.2.3: Planning of second set of applications (M19) [cancelled]
  • MS4.2.4: Planning of first set of applications (M25) [rescheduled M27]
  • MS4.2.5: Planning of second set of applications (M31)[rescheduled M32]

T4.3 (M1-36): Dissemination.

Milestones:

  • MS4.3.1: Set-up of the project Web server (M2)
  • MS4.3.2: Newsletters (quarterly)
  • MS4.3.3: Developers Workshops (M6, M18, M24)
  • MS4.3.4: Dissemination Workshops (M11, M23, M35)
Last update 13/06/2016
Universidad Autónoma de Madrid, Escuela Politécnica Superior
Video Processing and Understanding Lab