Network Computation in Neural Systems · 2003 · 796 citations · 59 references
ComputingComputer ScienceVision ScienceComputer VisionMachine VisionVisual PerceptionFeature DescriptionFeature DetectionImage ClassificationObject DetectionObject RecognitionScene UnderstandingNatural Image StatisticsNatural Scene StatisticsObject CategorizationObject ClassificationObject LocalizationObject-scene RelationsScene ScaleVisual Categorization
Second‑order statistics correlate with image categories, scene scale, and objects. The study investigates the statistical properties of natural images across categories and their relevance for scene and object categorization. The authors propose a feedforward approach to scene categorization that supplies top‑down and contextual cues early in visual processing. Visual categorization using low‑level features alone improves object localization and identification, and simple image statistics can predict object presence or absence before image exploration.
AbstractIn this paper we study the statistical properties of natural images belonging to different categories and their relevance for scene and object categorization tasks. We discuss how second-order statistics are correlated with image categories, scene scale and objects. We propose how scene categorization could be computed in a feedforward manner in order to provide top-down and contextual information very early in the visual processing chain. Results show how visual categorization based directly on low-level features, without grouping or segmentation stages, can benefit object localization and identification. We show how simple image statistics can be used to predict the presence and absence of objects in the scene before exploring the image.
59
Matthew Turk, Alex Pentland · Journal of Cognitive Neuroscience · 1991 · 13.8K citations · Full text
Recognition-by-components: A theory of human image understanding.
Irving Biederman · Psychological Review · 1987 · 5.6K citations