Paving the Way for Next-Gen Architecture With 3D Point Cloud Annotation

3D Point Cloud Annotation

Mindy Support played a pivotal role by delivering meticulous 3D Point Cloud data annotation services tailored to a broad spectrum of building designs. This comprehensive assistance was specifically designed to facilitate the client in bringing to fruition a project that had been commissioned by the architectural division of the institution. The adept annotation services provided by Mindy Support not only ensured precision in data representation but also contributed significantly to the seamless execution of the architectural vision embedded in the project.

The client’s architectural division commissioned a project to create an object recognition AI model that can work with objects from the real world. This meant that it needed to be able to recognize various elements of the construction design, such as walls, windows, doors, and many other things. The AI model would be trained on 3D Point Cloud images, and the client was looking for a trusted and reliable partner who had extensive experience in 3D Point Cloud annotation to help them actualize the project. 

The size of the data was quite large, consisting of more than 5000 images. In addition to this, the client was looking for a partner that could help them meet tight deadlines to keep the project on track and, therefore, requested that the entire dataset be annotated within 3 months. 

The client evaluated the portfolio of works of many data annotation vendors and shortlisted the candidates for the final round, one of which was Mindy Support. All the companies that reached this final stage of the tender process were invited to complete a pilot project. After evaluating the work of all the participants, the client was very impressed by Mindy Support’s attention to detail and overall quality of work and awarded us the project. 

Mindy Support assembled a team of 12 data annotators to complete all the 3D Point Cloud annotation work. All the team members we hired had many years of experience with this type of data annotation and were able to complete their work with minimal supervision required from the client. Once the team was assembled, we started off by analyzing all possible cloud points and dividing them into subgroups. In many cases, we had to annotate entire floors that could be more than a million points and clearly evaluate them. Even with such a large volume of data points, we maintained a high level of attention to detail while providing continuous delivery for the client. 

Thanks to the extensive experience and tech-savvy of our team members, we were able to complete the entire project within the timeframe specified by the client. We also managed to achieve a 98%+ quality score, which was particularly impressive given the scope of the dataset and the attention to detail required.

• 5000 images annotated within 3 months 
• Millions of data points placed with a 98%+ quality score
• 12 highly experienced team members onboarded