"Blind" visual inference by composition

Abstract “Blind” visual inference can often be performed by exploiting the internal redundancy inside a single visual datum (whether an image or a video). The strong recurrence of patches inside a single image/video provides a powerful data-specific priorfor solving complex visual tasks in a “blind” manner. The term “blind” is used here with a double meaning: (i) Blind in the sense that we can make sophisticated inferences about things we have never seen before, in a totally unsupervised way, with no prior examples or training; and (ii) Blind in the sense that we can solve complex Inverse-Problems, even when the forward degradation model is unknown. This paper briefly reviews this approach and its applicability to a variety of vision problems, ranging from low-level to high-level, including: (1) “Blind Optics” – recover optical properties of the unknown sensor, or optical properties of the unknown environment. This in turn gives rise to Blind-Deblurrimg, Blind-Dehazing, and more. (2) Segmentation of unconstrained videos and images. (3) Detection of complex objects and actions (with no prior examples or training).