The generic viewpoint assumption in a framework for visual perception

By William T. Freeman. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Group invariance principles for causal generative models (2017, background): “ in Fig. 1b, that would give the “illusion” that a yellow square is occluded. Such considerations have led vision scientists to formulate a generic viewpoint assumption in order to perform inference [Freeman, 1994]. Such scene understanding problem can be considered as a causal inference problem, as”
Learning to Reconstruct Shapes from Unseen Classes (2018, background): “1 The Effect of Viewpoints on Generalization The generic viewpoint assumption states that the observer is not in a special position relative to the object [Freeman, 1994].”
Bayesian models of object perception. (2003, background): “S2 pðS1; S2; I1Þ Discounting illumination by integrating it out can reduce ambiguity with regards to the location or shape of an object [34,35].”
A Bayesian framework for concept learning (1999, background): “…(1997), and has roots in the study of genericity phenom-ena in vision (Witkin & Tenenbaum, 1983; Koenderink, 1979; Binford, 1981; Lowe,55 1985; Freeman, 1994), Barlow's theories of cerebral cortex (Barlow, 1985, 1996), andGarner's information theoretic analysis of category structure (Garner,…”
Scene Summarization for Online Image Collections (2007, methodology): “The most popular criteria in previous work on canonical views are likelihood ([16], [28]) and orthogonality ([12], [17]).”
View-based 3-D object recognition using shock graphs (2002, methodology): “requires that we explore other alternatives, such as assigning probabilities of occurrence to the views and to the objects [ 17 ].”
Learning Low-Level Vision (1999, unknown): “The shape explanation for the crescents image requires nongeneric alignment of the assumed lighting direction (from the left) with the inferred shape (Freeman, 1994).”
Why Do Line Drawings Work? A Realism Hypothesis (2020, background): “Under the generic position and lighting assumption, as derived by Freeman (1994), there are three possible interpretations for any given class of strokes: • The strokes approximate shading, such as for smooth contours, earlier.”
The Innate Mind: Structure and Contents (2005, background): “…simple examples, similar judgments of probability and coincidence can explain many (or even most) other well-known aspects of visual processing (for extended discussion, see Hoffman, 1998; for specific discussions of ‘‘generic viewpoint’’ assumptions in vision, see Biederman, 1987; Freeman, 1996).”
Object perception as Bayesian inference. (2004, background): “The task itself can serve to reduce ambiguity by discounting the confounding variable (Freeman 1994,Kersten 1999, Schrater & Kersten 2000, Geisler & Kersten 2002).”
Learning about Canonical Views from Internet Image Collections (2012, background): “…concluded with the rather vague explanation that: “Canonical views appear to provide the perceiver with what might be called the most diagnostic information about the object: the information that best discriminates it from other objects, derived from the views from which it is most often seen” [3].”
Part-Based Representations of Visual Shape and Implications for Visual Cognition (2001, background): “Now, given a silhouette produced by a 3D shape of unknown geometry, the principle of genericity (e.g., Freeman, 1994) assigns high probability to those 3D interpretations in which the shape is about as deep as it is wide in the image.”