Citadel AI provides i-PRO, a global leader in sensing technologies for intelligent surveillance, with AI validation tools

Citadel AI (HQ: Shibuya, Tokyo; CEO: Hironori “Rick” Kobayashi), a leader in trustworthy AI technology, is pleased to announce that i-PRO Co., Ltd. (HQ: Minato-ku, Tokyo, President: Shohei Ozaki, hereinafter referred to as “i-PRO”), a global leader of advanced sensing technologies in the fields of Intelligent Surveillance, Public Safety and Industrial/Medical Imaging, is using Citadel Lens to improve the reliability of its AI systems. 

Challenges of AI for Security Monitoring

In the fields of security monitoring and public safety, the role of AI is more important than ever. AI efficiently analyzes large volumes of data collected from numerous security cameras, enabling real-time detection and identification of individuals, vehicles, and various objects. This capability allows for the swift detection of suspicious objects or activities, facilitating the rapid response and handling of potential risks.

However, AI models in security monitoring face challenges in accurately detecting and identifying multiple objects in diverse environments. Factors such as object size, angle, brightness, and color vary significantly depending on the camera setup and environmental conditions. Training AI to accurately pick out and identify target objects among numerous others in unexpected situations is highly challenging and requires considerable time and resources.

If these AI systems are trained poorly, the risks extend beyond technical quality issues to potential risks to human life, safety, corporate brand, and compliance with regulations and standards.

Enhancing the Reliability and Quality of Security Monitoring AI with Citadel Lens

Object detection AI models commonly experience reliability issues such as missed detections, false detections, bounding box misalignment, and misclassification. Citadel Lens automatically identifies and visualizes these issues, allowing for easy analysis of each target object based on metrics such as IOU, Recall, and Precision.

A unique feature of Citadel Lens is its ability to select individual objects and generate prediction heatmaps, known as Local Explainability. This enables visualization of the prediction rationale for each target object, making it easy to analyze why issues like false detections occur and utilize the insights for quality improvement.

Additionally, Citadel Lens automatically generates synthetic test data based on real-world environmental conditions, such as changes in camera hardware, environment, and positions of objects, allowing customers to immediately identify and improve areas of low model performance. 

The Lens Bias Finder automatically discovers weak areas of the model – slices of data where a model has abnormally low performance. This makes it easy to slice AI performance across dimensions such as the location of the image, the data annotator, the brightness of the environment, and the focus of the camera.

Collaboration between i-PRO and Citadel AI

Through the AI validation technology inside Citadel Lens, Citadel AI aims to further enhance the reliability and quality of i-PRO’s intelligent surveillance AI applications. Takeshi Shimada, from i-PRO’s Technology Engineering UX & AI Development AI Application Team, expressed, “Our vision is to continuously create new value contributing to a safer and more peaceful world through the provision of products and services. By collaborating with Citadel AI, we aim to conduct risk analysis and evaluation of AI applications using various accuracy metrics, achieve high reliability, and contribute to the realization of a safer and more peaceful world.” Citadel AI is committed to making the world’s AI systems more reliable.

About Citadel AI

Citadel AI provides software products that test and monitor the quality of AI systems. Our technology helps organizations minimize AI reliability risks and maximize AI performance from research to deployment. Citadel AI’s technology is built from our team’s first-hand experience deploying high-risk AI systems at world-leading companies such as Google Brain, Waymo, Toyota, and more. 

About i-PRO

i-PRO is a global leader of advanced sensing technologies in the fields of Intelligent Surveillance, Public Safety, and Industrial/Medical Imaging. Established in 2019, i-PRO was built on a legacy of over 60 years of innovation with Panasonic.

The company’s products, software and services extend human senses to capture moments of truth with innovations that inform and protect. In order to help create a safer world, i-PRO supports the work of professionals who protect and save lives.