Solution for Privacy Protection
Case study
Solution for Privacy Protection: Implementing AI-driven recognition and anonymization of people and vehicles in geospatial data
To comply with GDPR regulations, we developed an automated pipeline for recognizing and anonymizing individuals and vehicles in high-resolution images and videos collected by Field Group.
Who’s our partner?
Field Group
By combining software, sensors, and autonomous technologies, we put data to work to boost efficiency and reliability within the infrastructure, construction, environment, and public sectors.
source: www.field.group
We believe that the success of any project is based not only on task completion but, above all, on building relationships founded on trust, mutual understanding, and clear communication.
From the very beginning, we establish a routine of regular and transparent communication, set clear expectations, and promptly respond to changes in scope or emerging issues. This proactive approach ensures that the project stays on track and no time is wasted.
As part of their geospatial operations, Field Group collects and processes large amounts of high-resolution visual data.
As part of their geospatial operations, Field Group collects and processes large amounts of high-resolution visual data. However, GDPR regulations require companies to respect data privacy, making it necessary to automate the detection and anonymization of people and vehicles appearing in the footage.
Challenges included:
Early-stage anonymization ensuring privacy protection as soon as possible in the data collection or processing pipeline.
Maintaining high data quality anonymization should not degrade the overall usability of the dataset.
Efficiency and scalability the process should work seamlessly without delaying production workflows.
What is our proposition?
An AI-powered machine learning pipeline for automated anonymization of geospatial data
To meet the client's requirements, we developed and trained a software and machine learning processing pipeline for fast and accurate recognition and anonymization of individuals and vehicles.
Key features of our solution:
Accurately identifying people and vehicles in collected images and videos.
Precisely masking detected objects without affecting the quality of geospatial data.
Leveraging GPU-accelerated computing to ensure real-time processing.
Embedding the anonymization process into Field Group’s existing workflows.
By automating anonymization at an early stage, data privacy compliance is ensured without delaying production or reducing data quality.
The tech stack
Our solution leverages state-of-the-art technologies in computer vision, deep learning, and image processing, ensuring efficient and accurate anonymization of geospatial data. We utilized C++ and Python as core programming languages, integrating powerful libraries such as PyTorch, MMDetection, and OpenCV to enhance model performance.
To optimize processing, we employed TensorRT and Docker for infrastructure management, enabling scalable and high-speed computations. The system is built around advanced image segmentation and object detection techniques, utilizing cutting-edge neural networks like Mask R-CNN, Segment Anything, and YOLO to deliver precise and reliable anonymization.
Our expertise in AI, deep learning, and computer vision allowed us to develop a high-performance solution tailored to Field Group’s needs.
The anonymization pipeline integrates seamlessly into their existing geospatial data collection and processing workflows, ensuring compliance while maintaining efficiency.
ISO 27001 Certified Partner
We are excited to announce a key milestone in our commitment to high data security and information management standards: achieving the globally recognized ISO 27001 certification for our information security management system (ISMS).
Working with Codeflyers has been a game-changer for us! Their innovative approach and unparalleled technical expertise have propelled our project to new heights. We're thrilled with the progress and their proactive attitude.
Christoffer Apneseth
CTO at Field
Code Flyers Sp. z o.o.
NIP 7252322818
Piotrkowska 21/17
90-406 Łódź, Poland
Email: hi@codeflyers.com
Follow us
We believe that business and IT needs must be solved in a new and more future-oriented way. Needs must be realized faster and the solutions must create more value to maintain competitive advantage in a digital world that moves at a fast pace.