An autonomous micro-mobility startup that is developing an autonomous delivery vehicle to deliver small packages came to us with a requirement to annotate images. Their vehicles are designed to automatically detect obstacles, recognize road and traffic signs during the delivery of the package to the right customer at the expected delivery time.
Our client had hundreds of hours of sidewalk video split into individual frames for annotation. To achieve high levels of precision and accurate results, our client's data science team has to feed their neural networks with an immense amount of high-quality ground truth data. They then laboriously annotated every obstacle, road, and traffic sign in each frame. By doing so, they were able to map the coordinates of a person or product on an image. It would then indicate to the algorithm that an area on the image is of a person or a product. This is an important but extremely time-consuming part of their data science team’s responsibilities. The process can take up to 70% of their time and often takes away the team’s attention from other important tasks. Annotating the training data in-house was cost-ineffective and tiresome for the team.
Together, we discussed which solution would best fit their needs. Float needed accurate annotations to get optimal performance and quick turnaround time. We began with understanding the context in which the delivery vehicle would operate and started our annotation process.
They wanted their data set to be annotated as per the Cityscape dataset classification. The Cityscapes Dataset focuses on semantic understanding of urban street scenes and after a quick analysis of the dataset, we decided that we had to annotate every image with about 15 classes. They insisted in particular on the annotation precision. Even a few small errors can generate too much noise and dramatically reduce the algorithm prediction score. At DataClap, our in-house team of trained annotators allows us to work quickly and adapt to the needs of our clients.
Within a couple of weeks, we were able to do complete semantic segmentation of over 20,000 images with roughly 100,000 classifications across all images whilst providing high-quality results and excellent accuracy.
Here's what the founder of the startup had to say about our services,
"We have previously used a number of US based services for our image annotation tasks but none were able to provide the personalized attention to detail like DataClap did."
Our data annotation efforts helped our client implement a smart city project with challenging requirements.
We helped a sports analytics leader in developing AI models for sports analytics.
Important things to look for when choosing a data annotation partner
Case study on how we used Human in the Loop for data curation
Case study on how we used Human in the Loop for a document processing use case
Check how we helped a client in increasing the performance of their models.
Contact us with your requirements and we will set up a team to wok on your free pilot project. No commitments on your side.