As I said originally, it depends on what the actual structure of the data is. My assumption was that the data is a 2D image. By default, that is rendered as a slice, and as such the point data is rendered as pixel blocks (so it rather looks like cell data). (Lighting is also turned off for this instance, which is why I did not mention lighting.) If my assumption was right Cell Data to Point Data wouldn’t do anything because the data is already point data.
But if my assumption was wrong and the field is actually cell data, then my suggestion of resampling wouldn’t work at all and Cell Data to Point Data is the way to go.