Optimised brush_stroke_mask function for improved performance

[simdanonline]

This pull request optimises the brush_stroke_mask function to improve performance. These changes result in a significant performance boost, making the function more efficient and scalable. This optimisation is particularly important for large-scale image processing tasks, where every millisecond counts.

Changes:

• brush_stroke_mask function optimised for performance

• Vectorised operations used instead of loops

• TensorFlow's built-in image processing functions utilised

• Simplified logic and reduced unnecessary operations

• Improved code readability

[zuruoke]

thank you for your input :)