LectureSight is awesome and getting better!?

Feb 14, 2018Download as PPTX, PDF1 like241 views

LectureSight (http://lecturesight.org) is an open source OSGI application that uses OpenCL to analyze a video stream in real-time and track the presenter by controlling a PTZ camera. This presentation from the Opencast 2018 Conference in Vienna presents recent improvements to LectureSight from the University of Manchester and University of Cape Town.

||
Stephen Marquard, University of Cape Town
Stuart Phillipson, University of Manchester
Vienna, February 2018
14/02/2018 1
LectureSight is awesome and getting better!

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 2
Overview
LectureSight at University of Manchester
LectureSight at University of Cape Town
What��s new in LectureSight:
? More committers
? New documentation
? Improved UI
? Pan & Tilt operator
? Easier calibration
? Better tracking
? Metrics
? Galicaster plugin
? Dashboard

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 3
LectureSight at University of Manchester
(Demo video from Manchester)

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 4
LectureSight at University of Cape Town
Around 18 larger venues equipped, in daily use
Axis 5915 PTZ Cameras
Logitech and Raspberry Pi overview cameras
Some difficulties:
? Overview cameras become misaligned
? Overview cameras not positioned optimally
? Calibration sometimes not accurate
? ��Distracted�� tracking (presenter not in frame)
? Poor performance in very low light conditions
? USB/CAT5 extender and RPi camera issues
? Some venues need tilting as well as pan

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 5
More committers
? Ben Wulff (creator, now at Fraunhofer)
? James Perrin (UoM)
? Stephen Marquard (UCT)
We will review and merge your pull request promptly ?

|| 14/02/2018 6
New documentation: docs.lecturesight.org

|| 14/02/2018 7
Improved UI
PTZ Camera Control
now shows:
? Overview image as
background
? PTZ Camera field of
view
? Position of the
tracked object, target
camera position and
actual camera
position
Much easier to verify
that calibration and
configuration is correct.

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 8
Pan & Tilt Operator
Contributed by Manchester
Camera now tracks vertically as well as horizontally

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 10

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 11
Better tracking
Tracking persistence has been improved, so the tracker remains
on an object for longer, and is therefore less distracted by transient
movements elsewhere in the image (for example on projector
screens, or encoding artefacts).
Preference is given to tracked objects which have moved some
distance since initial detection (real people)

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 12
Metrics
LectureSight will
create a metrics
JSON file for each
session (from start to
stop of recording, if
using the CA��s
recording calendar)
Use metrics to
identify poor tracking
performance.
Use the Galicaster
plugin to attach the
metrics to the OC
mediapackage.

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 13
Galicaster plugin
Activates and deactivates LectureSight when recordings are
started and stopped manually in Galicaster
Attaches the LectureSight metrics file to the Opencast
mediapackage.

|| 14/02/2018First name Surname (edit via ��Insert�� > ��Header & Footer��) 15
Get LectureSight
http://docs.lecturesight.org/
http://bitbucket.org/bwulff/lecturesight/ (latest: 0.3-sprint)
https://opencast.jira.com/projects/LS/issues
Dashboard:
https://bitbucket.org/cilt/lecturesight-overview
Galicaster plugin:
https://bitbucket.org/cilt/matterhorn_ansible/raw/master/templates/lecturesight.py

The document discusses updates to Microsoft Azure products including: 1. New larger VM sizes in the B-series with increased CPU usage flexibility and improved performance calculations. 2. Updates to Azure Lab Services including support for GPUs, connecting to on-premises resources, and saving VM images for reuse. 3. Pricing details for Azure Lab Services based on instance size and lab units billed hourly.

KaTe AMQP Adapter for SAP PI/PO and Azure Service BusKate_RESTful

SIMCLOUD: Running Operational Simulators in the CloudFinmeccanica

This document discusses running operational spacecraft simulators in the cloud. It presents SIMCLOUD, a concept that uses cloud computing to deploy and run simulators. The SIMCLOUD system includes a cloud agent, cloud daemons, and a cloud management API to launch simulators on virtual machines in the cloud. Testing showed running simulators virtually incurred 10-30% overhead compared to physical machines. The authors conclude virtualization and cloud concepts can provide benefits like simplified use and elastic resources for simulator users.

AWS Partner Webcast - Create Instant Replicas of On-premises SAP LandscapesAmazon Web Services

You need to clone existing SAP landscapes for rapid prototyping of functionality, testing upgrades and patches, and for development sandboxes and POCs. But that can sometimes be complex and time consuming. Learn how Intrigo Systems, an SAP implementation organization, partners with Ravello to instantly create replicas of existing on-premises VMware-based SAP application environments on AWS, without changes to the VMs or networking. Webinar topics include: -Live demo showing how to save SAP blueprints and spin up your own SAP Landscapes on AWS via self-service. -How to create instant replicas of your on-premises VMware-based environments in AWS. -How to clone your on-premises SAP Landscapes in AWS for rapid prototyping, development and testing.

Juniper Unmanned AU PresentationJeff Cozart

This document summarizes a presentation on using unmanned aircraft systems (UAS) to calculate stockpile volumes. The key points are: 1. UAS can be used to acquire aerial imagery of stockpiles, which is then processed into 3D point clouds and analyzed in software like Reality Capture and Civil3D to calculate stockpile volumes. 2. A case study is presented where a UAS was used to map a wood chip stockpile, calculating the volume within 0.3% of a mobile scanning solution, but at a lower cost and faster acquisition time. 3. The presentation reviews the UAS data acquisition workflow, photogrammetric processing, point cloud generation, preprocessing in Recap and Civil

Helm your way with KubernetesAna-Maria Mihalceanu

Ana Maria Mihalceanu presented on using Helm to manage Kubernetes configurations. Helm allows deploying many Kubernetes configurations as a single application and parametrizing configurations for multiple environments. It was demonstrated using a chart per service approach and a master chart approach. When using Helm, it is important to check nested values, avoid hardcoding secrets, manage environments through values files, and mind chart organization.

PresentationDhanasekaran Nagarajan

The document discusses Infotech's automated testing services including eCATT testing, SAP testing capabilities, and case studies using tools like Mercury, QTP and Load Runner. It provides an overview of Infotech's 60 member SAP team and experience developing test scripts and procedures for clients' business processes. Sample projects are described for United Technologies Research Center involving ABAP development and eCATT, and for Metro Group Information Technology developing automation test scripts using QTP.

Comparative Analysis of IT Monitoring Toolsapprize360

This document provides a comparative analysis of IT monitoring platforms from CA Technologies, SolarWinds, IBM, and Nagios. It finds that CA Nimsoft Monitor version 7.5 had the fastest time to begin monitoring devices, being able to monitor 100 devices within 61 minutes of downloading and installing the software. In comparison, it took over 3 hours for Nagios XI and over 2 hours for some SolarWinds products to complete the same tasks. The analysis concludes that CA Nimsoft Monitor offers a single, unified monitoring solution while competitors require multiple individual products, and that CA Nimsoft Monitor's installation and configuration process is much easier and faster than the alternatives.

Wavefront-by-VMware-April-2019Anil Gupta (AJ) - vExpert

Wavefront provides a unified platform for real-time monitoring and observability across cloud infrastructure, applications, and services. It collects and analyzes time series metrics and distributed traces at massive scale to help users detect issues faster, optimize performance, and gain interactive insights. Key features include out-of-the-box integrations, real-time metrics ingestion, advanced analytics, and use cases for monitoring microservices, Kubernetes, containers, and public clouds.

Legacy Renewal of Central Framework in the EnterpriseAnatole Tresch

Monitor your spring boot application performances using inspectit apmAlain Lompo

In correlation with the fast paced technological evolution, software applications have also evolved to become more complex, more distributed and more dynamic. Cloud computing has pushed this evolution a step further. Therefore it is now more difficult to properly monitor application performances and ensure the end user that he will get the high-quality experience he is expecting. Luckily great tools such as InspectIT APM are taking care of the problem efficiently. In this presentation I will introduce you to inspectIt APM and then I will show you how to integrate it to your Spring boot application and how to diagnose, analyse and monitor your application. You will become familiar with the use of inspectIt three main components: the agent, the server and the user interface. You will also learn how to analyse user request from the invocation sequence and detect the root causes of potential problems. You will finally see how easy it is to perform real user monitoring with inspectIt in a production like environment.

IRJET- Video Based Traffic Sign Detection by Scale Based Frame Fusion TechniqueIRJET Journal

This document proposes a video-based framework for detecting, tracking, and recognizing traffic signs from a camera mounted on a moving vehicle. The framework aims to improve detection accuracy by utilizing information from spatially distributed traffic signs. It uses an offline detector initially, and if that fails, employs an online detector to update the motion model. Tracking is done using a Kalman filter to predict sign locations. The detections and tracks are fused to obtain a stable classification result. The proposed approach aims to handle non-stationary environments and improve location accuracy compared to single-frame detection methods.

DevOps monitoring: Best Practices using OpenShift combined with Icinga & Big ...Icinga

Talk by Marco Bizzantino: In my talk I will demonstrate how monitoring works with the DevOps approach. Within an organisation, different people with different roles need to have a complete view over the whole infrastructure. One of the greatest benefits of Icinga is its ability to integrate with various tools and modules to satisfy the company��s requests. An example is OpenShift monitoring. I will point out the most important parts of monitoring OpenShift infrastructure while maintaining it��s complexity. In my approach I will show how we combine Icinga, Prometheus, Grafana and Elasticsearch for an overall OpenShift monitoring solution. Additional examples will show the benefits of using Icinga��s Business Process modelling to add more value to the result.

Survey Paper for Different Video Stabilization TechniquesIRJET Journal

This document summarizes and compares three video stabilization techniques: Scale Invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF), and block-based methods. SIFT extracts distinctive keypoints from videos that are invariant to scale and rotation, but it is computationally slow. SURF is faster than SIFT and also extracts robust features. Block-based methods partition video frames into macroblocks and estimate motion between frames by block matching, using metrics like mean absolute difference. It has lower complexity than SIFT and SURF but provides good stability for video stabilization. The document analyzes the performance of these techniques and their application in video stabilization.

SiriusCon2016 - Modelling Spacecraft On-board Software with SiriusObeo

>> These slides were presented at SiriusCon Paris 2016, on November 15th by Andreas Jung (European Space Agency) The European Space Agency, together with industry, has lead an analysis into the issues faced by spacecraft software developers now and in the future, considering several aspects as for example raising complexity of the software, shorter development life cycles, etc. The analysis resulted in the development of an On-board Software Reference Architecture (OSRA) founded on the principles of component-based software engineering (CBSE) and strong separation of concerns. A dedicated Domain Specific Language for the component model was developed, called Space Component Model (SCM), to allow the precise definition with clear semantical meaning, in particular considering the domain specific elements like observability and commandability of spacecrafts via Telemetry and Telecommand. The SCM was implemented as a meta-model in ecore. The R&D activity that have developed the OSRA and the SCM have also prototyped a graphical editor to experiment and test the complete approach, from modelling down to code generation for the target. The original prototype of the graphical editor was based on Eclipse and Obeo Designer, which allowed very quick and simple prototyping of a graphical editor. Following the R&D activities, it was clear that an improved version of the editor, in terms of usability, is needed. An improvement activity has been started with Obeo, using now the open source version of Obeo Designer, namely Sirius. The intention was also to push Obeo's technology further to evaluate it for applicability in a commercial tool. This talk will give a brief overview of the challenges of spacecraft software development, the needs for a graphical editor, present the results of the improvement activity, show the benefits of the Eclipse and Sirius frameworks and provide an overall evaluation.

Can We Make Maps from Videos? ~From AI Algorithm to Engineering for Continuou...DeNA

�ɥ饤�֥쥳�`��ӳ�� AI �ǽ��뤳�Ȥǵ؇��ɤ��ƥʥ󥹤�ͥ��Ȼ��뤿��ȡ��M�ߤˤĤ��ưk��ޤ�� Ȥʤ륳��ԥ�`��ӥ��g��ʵĤ��_�k��֧�� DeNA�ʤ�ǤΥ��󥸥˥��⤴�B�餷�ޤ��

ARTIST: a global approach to cloudify applications, OW2 Open Cloud Forum at C...Ocean Project

Surrogate Model-Based Reliability Analysis of Composite UAV Wing facilitation...Altair

Numerical simulation becomes increasingly strategic to design innovative products and to set up their manufacturing processes, reducing simultaneously development costs and time to market while increasing quality and reliability. To support this evolution, SILKAN develops a platform for the integration of various types of simulation software, named BUILDERTM. BUILDERTM is an efficient, innovative and scalable simulation-based platform designed to deal with the increasing use of complex numerical simulations applied to part design, system design or manufacturing processes. The principal objectives of this platform are to: Promote and structure the use of simulation Standardize, parameterize and automate simulation processes. Capture and re-use the best practices. Facilitate coupling between different simulation levels and tools. Improve collaboration across different project teams. Facilitate access to simulation means for the uninitiated. Accelerate design and production cycles. Democratize the use of optimization and reliability procedures and better control manufacturing processes and failure risks. An application example using BUILDERTM is addressed in this paper. It deals with the robust design of a composite UAV wing. The associated simulation workflow includes two principal steps. During the first step, Matlab is used to estimate aerodynamic loads applied to the wing when as a function of flight parameters: air flow speed, angle of attack of the wing and aileron deflection angles. A Design of Experiment (DoE) is built by varying the flight parameters in order to cover all the flight domain of the UAV. The aerodynamic loads thus obtained are then injected into OptiStruct to estimate Tsai-Wu failure criteria for the composite material. An efficient surrogate model is then built from the obtained Tasi-Wu criteria and covers the entire flight domain. Finally to conclude this first part, a failure probability , based on Tsai-Wu criteria, is estimated using the produced surrogate model. In the second step the following optimization problem is defined using some design variables of the wing (essentially thicknesses of composite layers of the wing): Wing Mass is calculated by Optistruct, and being evaluated using the step1. An evolutionary algorithm implemented into Dakota is used to perform this surrogate-model - based optimization. The set up, parameterization and automation of this complex simulation workflow is facilitated and achieved through the use of the BUILDERTM platform. The combination of different software at different levels of the workflow is also made accessible by the use of BUILDERTM. Speakers Samir Ben Chaabane, Numerical Simulation Manager for EMEA, SILKAN S.A

SYM iSpectrum Business CasesTodd Masters

The iSpectrum functions as multiple virtual spectrum analyzers installed on site, providing 24/7 remote monitoring capabilities. It continuously records spectrum data, enables real-time spectrum viewing, and sends automatic SNMP alarms. This improves maintenance and troubleshooting by allowing post-event analysis without on-site support. Revenue opportunities exist in selling monitoring access and professional services using iSpectrum's tools and alarm notifications.

Case Study: Learn How Expeditors Uses APM as Both a Technology and Process T...CA Technologies

Expeditors implemented CA APM as both a technology and process transformation initiative. They took a three-tiered approach focusing on technology, people, and process. On the technology side, they deployed CA APM across their pre-production and production environments. For people, they defined roles, provided training, and partnered with experts. For process, they incorporated APM into their SDLC and created a maintenance discipline. The implementation helped Expeditors gain visibility, improve performance, and empower teams across the organization.

Back to FME School - Day 3: Expanding FrontiersSafe Software

It's that time of year. The season is changing and FME 'school' is now in session! Join us for a series of 9 mini-talks to learn the latest tips for data transformation, see live demos, and get your FME questions answered. Registration gives you access for all three days �� sign up now to tune in to the talks you're most interested in. Course Schedule - 40 Minute Talks: Expanding Frontiers - Day 3 - Remote sensing providers - Web and business data into gold - New + Upcoming in FME Cloud

Securescape IntroductionDaden Limited

Hands-On Lab: Let's Build an ITSM DashboardCA Technologies

Having the right information at your fingertips at exactly the right moment to get that project approved is critical. Static dashboards are no longer acceptable. This presentation shows how to build an ITSM dashboard, then change it is seconds, seeing the power and value that advanced reporting and dashboards bring to ITSM. Learn more about CA Service Managment solutions at http://bit.ly/1wEnPhz

Hidden Camera 3 APIs in Android 4.4 (KitKat)Balwinder Kaur

The document provides an overview of the new Camera 3 framework introduced in Android 4.3 and 4.4. It aims to expand the camera subsystem to support computational photography use cases like HDR, panoramic stitching, and multi-shot processing. The new framework models the camera as a pipeline that applications can control through fine-grained APIs to process image data and metadata in real-time. It allows applications to directly access image buffers and control camera settings per frame through a request-result model.

Cisco project ideasVIT University

FINken eYe: Visual SLAM-Based Position EstimationMichael Mera

Onion Architecture with S#arpGary Pedretti

The implementation of an Opt-Out Lecture Recording Policy at the University o...Stephen Marquard

The University of Cape Town (UCT) launched its Opencast lecture recording service in 2013. In 2017, a new Lecture Recording Policy was adopted, enabling opt-out recording for most undergraduate courses. This came into effect in July 2018, with an initial opt-out rate of around 40%. This presentation will describe the journal from a small pilot to a large-scale opt-out service, the campus response to the opt-out policy, and some of the debates which continue around the value and relevance of lecture recording on campus.

Orchestrating Self-Service Video Workflows with OpencastStephen Marquard

Presentation at the Open Apereo Conference, Montreal, June 2018. Video has come to play a central role in blended and online courses, and is increasingly being used for formative and summative assessment experiences. Educators and students very often want or need to create, edit and publish their own video, but without the difficulties that come along with DIY solutions or desktop-based editors. UCT has extended our Opencast video management system for self-service applications which include browser-based desktop recording, just-in-time recording in traditional classrooms, instructional videos in studio settings and students recording themselves in clinical settings for formative assessment. This presentation will outline design challenges and solutions for these different scenarios, and how Opencast workflows are being used to meet the needs of different user groups.

More Related Content

Similar to LectureSight is awesome and getting better!? (20)