Well it would be simpler to use a 1D array of PVectors; but you might still run into perfomance issues...  The beauty of using get(mouseX, mouseY) - even though get() is relatively slow compared to accessing the pixel array - is that you're only checking against a single point each frame; rather than checking the mouse location against "thousands of points".

In practical terms though it does depend on what the OP wants - e.g. is the response to being on a point the same for all points or do different points need to trigger a different response?  There are still ways around that using the get() approach - e.g. you could assign different behaviours to different colours and check the colour value on get(mouseX, mouseY) and respond accordingly.  If you want the display to present all points as the same colour you could use a separate buffer image with coloured points...