Best Practices & Use Cases

welcome to the practical guide to maskit while our other sections explain the tools, this section shows you how to use them to solve real world problems learn the best ways to approach common tasks and see how maskit handles even the most challenging scenarios articles in this section best practice choosing the right anonymization level best practice efficient high volume processing use case anonymizing security & traffic camera footage use case preparing images for real estate listings use case handling complex scenarios (crowds, reflections & bad weather) article 1 best practice choosing the right anonymization level one of the key questions in data privacy is how much information to anonymize maskit gives you the flexibility to choose, but which option is right for you? when to use "mask faces" this is the most common scenario use face only masking when your goal is to protect the identity of individuals while preserving the full context of their actions and attire ideal for market research photos, event photography, journalistic content where the person's activity is important when to use "mask whole persons" choose this for maximum privacy this method is essential when not just the face, but also unique clothing, tattoos, or body shape could lead to identification ideal for public space analytics (e g , measuring foot traffic), anonymizing bystanders in street level imagery for mapping services, or any scenario requiring the highest level of gdpr compliance the foundation of every choice accuracy regardless of which level you choose ("mask faces" or "mask whole persons"), the most critical factor is the accuracy of the detection maskit's high precision models ensure that whatever you choose to anonymize is found reliably, giving you peace of mind and forming the foundation of a robust compliance strategy article 2 best practice efficient high volume processing whether you're using the web ui or the api, these tips will help you process large volumes of files faster prepare your data before uploading a large batch or starting an api script, ensure your files are in a supported format and are not corrupted this prevents errors mid process use batch uploads in the web ui, always use the folder upload or multi file select feature for batches up to 100 files don't process high volumes one by one leverage asynchronous api our api is asynchronous for a reason submit your jobs, store the jobid, and use webhooks to be notified upon completion polling the status endpoint is also efficient if you check at reasonable intervals (e g , every 5 10 seconds), not multiple times per second article 3 use case anonymizing security & traffic camera footage maskit video processing is not available yet in security and public space monitoring, accuracy is non negotiable a single missed face or license plate isn't a minor error; it can be a significant compliance breach maskit's high accuracy models are built to handle the complexities of video footage, providing automated results you can genuinely rely on the challenge security cameras often capture hundreds of faces and license plates per hour the video quality can be low, and angles can be challenging the maskit solution upload video upload your footage via the web ui or api select all detections ensure "mask faces," "mask humans," and "mask license plates" are all selected for comprehensive coverage process & download maskit processes the video frame by frame, applying consistent anonymization to all detected objects, even as they move result you get a fully compliant video file that can be used for security analysis or as evidence, without violating the privacy of uninvolved individuals article 4 use case preparing images for real estate listings when showcasing a property, you need to remove any personal data from the photos, which can be more than just people the challenge photos of home interiors often contain personal family photographs in frames, faces on tv screens, or even people in reflections the maskit solution upload interior photos upload your real estate pictures to the web ui use "mask faces" our model is trained to detect faces not only on people but also in pictures within pictures it will automatically find and anonymize the faces in the photo frames on the wall check for reflections our ai also detects faces in reflections on windows or glossy surfaces result you can quickly and easily create gdpr compliant property listings that protect the privacy of the current residents article 5 use case handling complex scenarios this is where a powerful ai model shows its strength maskit is designed to handle difficult, real world conditions with maximum accuracy where simpler tools might fail challenge 1 crowds & overlapping people our model can identify individuals even when they are partially obscured by others in a crowd challenge 2 reflections & glass detecting a face that is reflected in a car window or visible through a storefront requires an advanced model maskit is trained to handle these cases challenge 3 bad weather conditions faces and license plates seen through a rain streaked window are heavily distorted our system is designed to identify sensitive data even in these low visibility conditions