Abstract:
The article describes an algorithm for blurring image boundaries based on fuzzy sets. The application of this algorithm to the blurring of simple contours is shown: stripes, squares, circles. The possibility of modeling a wake jet on the surface of water left by a moving vessel is shown. The proposed approach allows you to create a dataset for training convolutional neural networks capable of detecting an object in a blurred and noisy image.
Fulltext:
PDF file (1207 kB)