Solution Brief AI Vision/Deep Learning Situational Monitoring

Solution Brief
AI Vision/Deep Learning
Situational Monitoring

Solution Brief AI Vision/Deep Learning Situational Monitoring

Vulcan AI Uses AI Vision and Deep Learning to Create Safer Workplaces The enterprise AI solution provider helps protect against workplace hazards and the spread of infectious diseases by extending inference and deep learning capabilities to the edge. “Intel® tools and technologies helped us offer customers reliable, outstanding performance on cost-efficient edge ...hardware. The optimization we achieved ensures that WorkSafe can rapidly identify hazards and help keep our customers’ workforces safe.” — Kamal Mannar, PhD Head of Applied Intelligence, Vulcan AI to 3 5x 1 throughput increase using the Intel® Distribution of OpenVINO™ toolkit See backup for workloads and configurations. Results may vary. For all companies, employees are critical assets. And keeping them safe is a top priority for business leaders. Many organizations are seeking new ways to protect their team and more effectively address workplace hazards and health concerns. With the right tools, they can help ensure the well-being of their workforce while also preventing downtime, maximizing productivity, and enhancing business results. To help businesses accomplish their workplace safety goals, Vulcan AI created WorkSafe, an innovative solution for identifying employee safety issues such as spills, near misses, slips, falls, trips, and health risks in near-real time. It’s a forward-thinking application of AI vision at the edge that constantly assesses video feeds to identify hazards and incidents. As issues are detected, the system prompts users to take preventive and corrective measures. The WorkSafe solution also monitors behavior such as social distancing and mask wearing to reduce the risk of infectious disease spread. When issues are identified, the system nudges employees through visual cues and notifications that prompt them to realign with proper protocols. Common incidents are identified almost instantaneously. Facilities managers and safety management officers gain visibility of risk hotspots alongside precise location insights. Alerts and information can be accessed at any time through a web-based portal. And managers receive daily risk assessment reports that highlight high-risk areas. WorkSafe leverages stereo vision and AI pipeline technology to detect hazards and compliance with workplace safety protocols for infectious diseases. The solution also leverages wearable AI to pinpoint safety risks that are not within view of cameras. Multisensory inputs provide more-complete coverage that enables early and accurate risk detection. Challenge: Capturing and analyzing workplace hazard data at the edge in real time Workplace accidents and hazards require immediate on-site intervention. Video of these events is sensitive and private, which requires the WorkSafe solution to process data at the edge. To keep pace, WorkSafe requires low latency and optimized performance from the edge hardware it runs on. The solution needs to perform excellently on cost-efficient CPU- and VPU-based configurations to keep customer expenses manageable. To enable the required inference and incident identification, Vulcan AI worked extensively to generate the required data sets. Faced with a lack of data, particularly in the industrial sector, the team used simulations to create a custom database of normal activities, hazards, near issues, and incidents. They were then able to create robust models for accurate detection. Solution Brief | Vulcan AI Uses AI Vision and Deep Learning to Create Safer Workplaces Video systems capture footage Edge hardware, optimized and tested via Intel® technologies, analyzes video data for workplace incidents System sends notifications to address issues and captures incidents within simplified dashboards Figure 1. The Vulcan AI WorkSafe solution uses video data from workspace cameras and CCTV systems Read the full Solution Brief AI Vision/Deep Learning Situational Monitoring.