An approach to automate time and motion analysis
Who is this presentation for?
- AI technologists and industry (manufacturing) specialists
Time and motion study is typically carried out in manufacturing shop floor and other organizations, primarily for performance optimization and to improve quality. Sundar Varadarajan and Peyman Behbahani explore a study that involves procedural measurement of time taken by human operators for completion of various tasks and subtasks in an operation, involving usually a long repetitive sequence of tasks. As an example, they focus on the measurement of task times in an assembly line operation, which is typically carried out using a time-keeping device such as a stopwatch or a video camera to record a repetitive sequence of subtasks in the operation. This process involves manual steps and is laborious and time-consuming. Given the rapid growth of digital videos, in application areas such as shop floors and factory assembly lines, Sundar and Peyman detail an alternative approach for automation of task and subtasks time measurement using video analytics involving computer vision (CV) and machine learning (ML), and they share their results in applying this technique for an assembly line operation in a shop floor.
You’ll discover insights and statistics about the state of time, motion, and human activity analytics for repetitive activities in the manufacturing shop floor environment using a combined structural similarity index (SSIM) and time series analysis. Image sequence processing, image sequence analysis, and visualization are the main subsystems of this algorithm. Furthermore, unlike deep learning models, which require several hours of annotated video, this model can work with very little training data to define the signatures of repetitive activities.
Using traditional image processing and similarity ranking, you can to identify and measure times for some simple tasks. When combined with activity signature learning and use of deep learning to identify objects and human activities, this hybrid approach can automate measurement of time for more complex tasks and operations. The algorithm can be easily integrated to any analytical system that has access to video footage of the shop floors either via a batch or with near-real-time processing. This has shown promising results for the sample assembly line operation and subtasks and can be further enhanced for a broader deployment, where lighting conditions, camera positions, occlusion possibilities, are all taken into consideration.
What you'll learn
- Discover a new approach for automation of action identification, time and motion study, automation using video analytics, CV, image processing, and ML
Sundar Varadarajan is a consulting partner on AI and ML at Wipro and plays an advisory role on edge AI and ML solutions. He’s an industry expert in the field of analytics, machine learning and AI, having ideated, architected and implemented innovative AI solutions across multiple industry verticals. Sundar can be reached at firstname.lastname@example.org.
Peyman Behbahani is a senior AI architect at Wipro, helping various industries on building real-world and large-scale AI applications in their businesses. He earned his PhD in electronic engineering at City, University of London in 2011. His main research and development interest is in AI, computer vision, mathematical modeling, and forecasting.
Leave a Comment or Question
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
Premier Diamond Sponsors
Premier Exhibitor Plus
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires