Interference is an optics term related to combination of waves. It can also appear in other wave describable phenomena like electron imagery. Usually one will se an interference pattern. In a larger scale moire patterns between periodic, partially translucent media laid over each other, or one in front of a mirrored image.
Interference phenomena may be noise if it is unwanted. However, the phenomenon can be utilized in highly efficient and sensitive measurement systems.
I bounce on Ander's comment: "noise" is just what disturbs the result you want to get, it has no absolute definition out of the user's aim and toolbox. It contains both organized and unorganized components. Interference is an organized component as it comes from a multipath context, while for example thermal noise on your photodetector is unorganized. To reduce noise, you have to take into account the particulars of each noise component: when the relation between the ideal image and the spurious component (as for interference) is explicit, a model can be applied to correct the raw image (image correction models work as well if the camera detector has moved during the exposure, or to compensate permanent defects of lenses etc.); for perfectly random noise, averaging works well. Unfortunately many noise components are in the grey zone between deterministic and random - still powerful adaptative algorithms for image resolution improvement can work even with no explicit model of the noise - and make in particular astronomers very happy...