Summary
In this chapter, we learned why ML models need annotated real data. At the same time, we explored some of the common issues in the annotation process. Our exploration has led us to a deeper understanding of real data collection and annotation issues, such as being a time-consuming process and subject to annotator errors. Additionally, we covered the limitations of real data for tasks such as optical flow and depth estimation. In the next chapter, we will look specifically at the privacy issues with real data.
In the following chapters of the book, we will continue our exciting journey to understand the limitations of real data and the promising solutions of synthetic data.
