If you are able to get more data on which the algorithm can train, that can help the algorithm to avoid overfitting by focusing on general patterns rather than on patterns specific to small data points. There are several cases where getting more labeled data could be a challenge.

There are techniques, such as data augmentation, that can be used to generate more training data in problems related to computer vision. Data augmentation is a technique where you can adjust the images slightly by performing different actions such as rotating, cropping, and generating more data. With enough domain understanding, you can create synthetic data too if capturing actual data is expensive. There are other ways that can help to avoid overfitting when you are unable to get more data. Let's look at them.