In a real-world setting, a model is expected to interact with streams of dynamic data to capture some behaviour of the model, whereas, in a laboratory setting, you have just sets of static datasets with no access to an existing data stream source. The need here is to simulate a dynamic data stream from the sets of static data sets.
Christopher Ufuoma Onova Your question is not clear so far if you want to test your algorithm on a dataset. then I suggest TUM-VI and TUM-RGBD for static as well as dynamic scenarios.
@Christopher Onova that would mean that you are deploying the model for use either as a pluggable web service or integrated as part of a web application where real-time data or images can be fed into the model for analysis. If you post a gory picture on Facebook there is an inbuilt ML model that analysis and classifies the picture then covers the image if it is positively identified as offensive. The users are then warned to uncover the picture at their discretion. So your models can be integrated into web applications to access real-time data.
- Badges
- Science topic
- Similar topics
- Computer Science
- Program