Learning Object Representations from Lighting Variations
نویسندگان
چکیده
Realistic representation of objects requires models which can synthesize the image of an object under all possible viewing conditions. We propose to learn these models from examples. Methods for learning surface geometry and albedo from one or more images under fixed posed and varying lighting conditions are described. Singular value decomposition (SVD) is used to determine shape, albedo, and lighting conditions up to an unknown 3×3 matrix, which is sufficient for recognition. The use of class-specific knowledge and the integrability constraint to determine this matrix is explored. We show that when the integrability constraint is applied to objects with varying albedo it leads to an ambiguity in depth estimation similar to the bas relief ambiguity. The integrability constraint, however, is useful for resolving ambiguities which arise in current photometric theories. Object Recognition Workshop. ECCV. 1996.
منابع مشابه
Bilinear models of natural images
Previous work on unsupervised learning has shown that it is possible to learn Gabor-like feature representations, similar to those employed in the primary visual cortex, from the statistics of natural images. However, such representations are still not readily suited for object recognition or other high-level visual tasks because they can change drastically as the image changes to due object mo...
متن کاملOn Nonparametric Guidance for Learning Autoencoder Representations
Unsupervised discovery of latent representations, in addition to being useful for density modeling, visualisation and exploratory data analysis, is also increasingly important for learning features relevant to discriminative tasks. Autoencoders, in particular, have proven to be an effective way to learn latent codes that reflect meaningful variations in data. A continuing challenge, however, is...
متن کاملInvariant visual object recognition: a model, with lighting invariance.
How are invariant representations of objects formed in the visual cortex? We describe a neurophysiological and computational approach which focusses on a feature hierarchy model in which invariant representations can be built by self-organizing learning based on the statistics of the visual input. The model can use temporal continuity in an associative synaptic learning rule with a short term m...
متن کاملLearning to Recognize Objects - Toward Automatic Calibration of Color Vision for Sony Robots
Color detection can be seriously affected by lighting conditions and other variations in the environment. The robot vision systems need to be recalibrated as lighting conditions change, otherwise they fail to recognize objects or classify them incorrectly. This paper describes experiments toward object recognition under different lightning conditions. We propose to train the vision system to re...
متن کاملHuman perception of local shape from shading under variable lighting
It is commonly believed that humans exhibit lighting invariant object recognition. Here we present psychophysical evidence to the contrary. We examine the perception of qualitative local shape from shading and find that significant variations occur as a function of the lighting. These variations have two logically distinct causes: first, there are inherent ambiguities between shape and shading ...
متن کامل