Construction Workplace Safety
Threat detection +35%; false alarms reduced ~40% on Tictag-annotated worksite footage
A safety video analytics company partnered with Tictag to label diverse worksite footage, raising AI threat-detection accuracy, cutting false alarms, and accelerating safety decision-making across construction sites.
SafetyTech, a reputable safety video analytics company, operates in various industries to provide advanced security and safety solutions. With a growing number of safety concerns and potential threats, SafetyTech aimed to enhance the accuracy and efficiency of their AI-powered safety video analytics system. Their objective was to deliver are liable and effective system that enables rapid response and risk reduction.
Objective
SafetyTech aimed to enhance the accuracy and efficiency of their AI-powered safety video analytics system. The primary objectives were:
- Improve threat and safety violation detection rates.
- Reduce false alarms and improve system reliability.
- Speed up the decision-making process for safety and security professionals.
Challenges
Inaccurate AI algorithms: SafetyTech's existing algorithms were not performing at the desired level of accuracy, impacting their ability to detect threats and safety violations effectively.
Insufficient labeled training data: SafetyTech lacked a sufficient amount of high-quality labeled training data required to train their AI system and improve its performance.
Time-consuming data annotation: Annotating safety and security video footage with precision and accuracy can be a labor-intensive and time-consuming task, requiring significant resources.
Maintaining annotation quality: Ensuring a high level of annotation accuracy throughout the process posed a challenge, as errors or inconsistencies could impact the reliability of the AI system.
Balancing dataset diversity: Curating a balanced and diverse dataset that covers various safety scenarios and environments was crucial to ensure the AI system's performance across different situations, posing a challenge in dataset curation.
