Privacy Policy: For Video Dataset Subjects
Welcome
Who Should Read This Policy?This Privacy Policy is relevant even if you are not a user or customer of Project Hafnia but are curious about why and how we process personal data for R&D purposes. We explain how we process personal data to innovate, generate new knowledge, and develop products as part of our work at (Milestone System A/S - “Project Hafnia”, “PH”, “we”, “Milestone”), including how we:
- Collect, use, store, delete, and share personal data.
- Maintain confidentiality and apply privacy-preserving techniques by default.
- Ensure personal data is not distributed to the public nor accessed by unauthorized people.
- Never identify, target, or profile individuals in our research activities.
The Role of Personal Data in Advancing R&D at Project Hafnia
At Project Hafnia, we utilize visuals of personal data to advance our Research and Development (R&D) initiatives. Personal data play a crucial role in enabling us to develop products, improve computer vision awareness, and create technologies that benefit individuals, society, and the environment.The data and information we collect for our research are, as a general rule, kept confidential and are never made publicly available. We systematically apply privacy-preserving techniques by default to safeguard individuals' data.
As a rule, PH does not identify individuals, take measures against them, or engage in profiling as part of our research activities.
We are committed to ethical AI development and take pride in the contributions of our partners who source and process data responsibly. To reinforce this commitment, we have established clear ethical guidelines for:
Contacts
Milestone Systems A/S
Company Registration No: 20341130
Banemarksvej 50
2605 Brøndby, Denmark
Tel.: +45 88 30 03 00
Datatilsynet (Danish Data Protection Agency)
Carl Jacobsensvej 35
DK-2500 Valby
Denmark
phone + 45 33 19 32 00
Processing Data to Advance R&D at Project Hafnia
- Set New Industry Standards: Establish robust datasets and end-to-end compliance frameworks to advance computer vision development responsibly.
- Build a Compliant Data Space: Develop an indexed, adaptable, and compliant space for model development.
- Analyze Privacy-Preserving Techniques: Assess the quality and robustness of privacy measures applied to images and videos.
- Innovate Curation and Annotation Flows: Build advanced methods for curation and labeling of data.
- Optimize Models: Improve the accuracy and performance of computer vision models.
- Benchmark Models: Evaluate computer vision models for accuracy, performance, and biases.
- Advance VLMs and Foundational Models: null
- Create new training and fine-tuning techniques.
- Enable and deploy Vision-Language Models (VLMs) and foundational models effectively.
- Develop New Use Cases: Advance computer vision.
- Generate New Knowledge and Insights: Expand innovation in computer vision.
At PH, we are committed to acquiring data exclusively through official channels under dedicated license agreements. We do not use web crawlers, data scraping, or data mining to expand our data acquisition processes.This approach ensures ethical and transparent data acquisition practices in alignment with our privacy commitments.
We obtain personal data from the following trusted sources:
- Cities: Data provided through partnerships and agreements with municipalities.
- Our Own Cameras: Data collected directly by PH cameras and available Milestone Group data.
- Resellers and Partners: Data sourced from Resellers of Milestone solutions and their associated Partners.
- Universities: Data sourced through academic collaborations.
- Public and Open Source Datasets: Data available through publicly accessible and open sources, where the license allows it.
- Visual Data (Personal Identifiable Attributes):
- Facial features of individuals.
- Body shapes and physical movements.
- Actions performed in the footage (e.g., waiting at a signal, walking, cycling).
- Identifiable clothing or accessories (e.g., distinctive uniforms, logos, brands, or reflective gear).
- Vehicle-Related Data:
- Vehicle license plates.
- Vehicle brand, model, color, or other distinctive visuals.
- Contextual Data:
- Exact or approximate location where the video was recorded.
- Exact or approximate date and time of the recording.
- Visually identifiable landmarks.
Our processing of personal data requires a legal basis, and we ensure that all activities involving personal data are conducted in compliance with applicable laws and regulations.
We process it because it is necessary for our legitimate interests in conducting scientific research within the core areas in which we operate.
- Develop and improve our offerings to benefit users globally.
- Gain insights into how our products and services are used.
- GDPR Article 6(1)(f): Processing necessary for legitimate interests.
- GDPR Article 9(2)(j): Processing for scientific research purposes.
- Section 10 of the Danish Data Protection Act.
- When legally required, we may share data with public authorities upon request.
- Suppliers and Service Providers may require access to certain data to support and advance our R&D efforts.
- These include entities such as consultants, law firms, and other service providers.
- Research Organizations that support our research and innovation efforts may require access to certain data.
- These include entities such as research labs, universities etc.
- Close collaboration partners may require access to certain data to advance research, innovation and product development.
- Personal data may be transferred to countries that the EU Commission has deemed to have an adequate level of data protection.
- For transfers to countries without an adequacy decision, we rely on Standard Contractual Causes approved by the EU Commission.
- We conduct thorough assessments of the processor and the processing activity to ensure compliance with legal requirements.
- All use of processors and storage of personal data in third countries is carefully evaluated, considering the scope of the transfer, the safeguard level, and compliance with Data Processing Agreements (DPAs).
- We evaluate the relevance of the data every 7 years.
- If the data is still relevant, we may extend the retention period by an additional 2 years.
- This evaluation process will repeat at the end of each extended retention period.
- Destroy the data, ensuring it is permanently and irreversibly deleted.
- Anonymize the data, provided current techniques guarantee that re-identification is impossible.
- Our partners adhere to the same retention periods. They may delete data earlier if required or at our request, ensuring consistency in data handling and protection.
- If you are not a user of our platform, due to the nature of the personal data we process, we do not have your contact information.here. (LINK NEEDED)
- If you are a user of our platform,here.