Limits...
Fragment-based learning of visual object categories in non-human primates.

Kromrey S, Maestri M, Hauffen K, Bart E, Hegdé J - PLoS ONE (2010)

Bottom Line: Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category.We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so.However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear.

View Article: PubMed Central - PubMed

Affiliation: Brain and Behavior Discovery Institute, Medical College of Georgia, Augusta, Georgia, USA.

ABSTRACT
When we perceive a visual object, we implicitly or explicitly associate it with an object category we know. Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category. We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so. However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear. Here we describe the methods by which we adapted the relevant human psychophysical methods to awake, behaving monkeys and replicated key previous psychophysical results. This establishes awake, behaving monkeys as a useful system for future neurophysiological studies not only of informative fragments in particular, but also of object categorization and category learning in general.

Show MeSH
Naturalistic shape classes generated by virtual phylogenesis (VP).(A) The VP algorithm for generating naturalistic shape classes. This algorithm simulates biological evolution, in that shape characteristics evolve as random heritable variations are differentially propagated through successive generations [12], [13]. Note that the differences between, as well as within, the categories arise spontaneously and randomly during VP, rather than as a result of externally imposed rules, including the fragment selection process or any other classification scheme. The bottom of the evolutionary cascade denotes the three shape classes used in many of the experiments in this study. See refs. [12], [13] for additional examples of shape classes. (B) Shape variations within and across classes X, Y and Z as visualized by a metric multi-dimensional scaling (MDS) plot. Each data point represents one object from a given class (inset). MDS plots the data points so as to cluster similar data points together and disperse dissimilar data points from each other, so to provide a principled representation of the relevant classes (for details, see refs. [48], [49]).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2991334&req=5

pone-0015444-g001: Naturalistic shape classes generated by virtual phylogenesis (VP).(A) The VP algorithm for generating naturalistic shape classes. This algorithm simulates biological evolution, in that shape characteristics evolve as random heritable variations are differentially propagated through successive generations [12], [13]. Note that the differences between, as well as within, the categories arise spontaneously and randomly during VP, rather than as a result of externally imposed rules, including the fragment selection process or any other classification scheme. The bottom of the evolutionary cascade denotes the three shape classes used in many of the experiments in this study. See refs. [12], [13] for additional examples of shape classes. (B) Shape variations within and across classes X, Y and Z as visualized by a metric multi-dimensional scaling (MDS) plot. Each data point represents one object from a given class (inset). MDS plots the data points so as to cluster similar data points together and disperse dissimilar data points from each other, so to provide a principled representation of the relevant classes (for details, see refs. [48], [49]).

Mentions: We created several dozen naturalistic object classes using the VP [12], [13] algorithm described previously (see Fig. 1 and Materials and Methods for details). The categories were such that no two objects, including objects within a given category, were exactly alike. Therefore, distinguishing among them required learning the relevant statistical properties of the objects and ignoring the irrelevant variations.


Fragment-based learning of visual object categories in non-human primates.

Kromrey S, Maestri M, Hauffen K, Bart E, Hegdé J - PLoS ONE (2010)

Naturalistic shape classes generated by virtual phylogenesis (VP).(A) The VP algorithm for generating naturalistic shape classes. This algorithm simulates biological evolution, in that shape characteristics evolve as random heritable variations are differentially propagated through successive generations [12], [13]. Note that the differences between, as well as within, the categories arise spontaneously and randomly during VP, rather than as a result of externally imposed rules, including the fragment selection process or any other classification scheme. The bottom of the evolutionary cascade denotes the three shape classes used in many of the experiments in this study. See refs. [12], [13] for additional examples of shape classes. (B) Shape variations within and across classes X, Y and Z as visualized by a metric multi-dimensional scaling (MDS) plot. Each data point represents one object from a given class (inset). MDS plots the data points so as to cluster similar data points together and disperse dissimilar data points from each other, so to provide a principled representation of the relevant classes (for details, see refs. [48], [49]).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2991334&req=5

pone-0015444-g001: Naturalistic shape classes generated by virtual phylogenesis (VP).(A) The VP algorithm for generating naturalistic shape classes. This algorithm simulates biological evolution, in that shape characteristics evolve as random heritable variations are differentially propagated through successive generations [12], [13]. Note that the differences between, as well as within, the categories arise spontaneously and randomly during VP, rather than as a result of externally imposed rules, including the fragment selection process or any other classification scheme. The bottom of the evolutionary cascade denotes the three shape classes used in many of the experiments in this study. See refs. [12], [13] for additional examples of shape classes. (B) Shape variations within and across classes X, Y and Z as visualized by a metric multi-dimensional scaling (MDS) plot. Each data point represents one object from a given class (inset). MDS plots the data points so as to cluster similar data points together and disperse dissimilar data points from each other, so to provide a principled representation of the relevant classes (for details, see refs. [48], [49]).
Mentions: We created several dozen naturalistic object classes using the VP [12], [13] algorithm described previously (see Fig. 1 and Materials and Methods for details). The categories were such that no two objects, including objects within a given category, were exactly alike. Therefore, distinguishing among them required learning the relevant statistical properties of the objects and ignoring the irrelevant variations.

Bottom Line: Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category.We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so.However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear.

View Article: PubMed Central - PubMed

Affiliation: Brain and Behavior Discovery Institute, Medical College of Georgia, Augusta, Georgia, USA.

ABSTRACT
When we perceive a visual object, we implicitly or explicitly associate it with an object category we know. Recent research has shown that the visual system can use local, informative image fragments of a given object, rather than the whole object, to classify it into a familiar category. We have previously reported, using human psychophysical studies, that when subjects learn new object categories using whole objects, they incidentally learn informative fragments, even when not required to do so. However, the neuronal mechanisms by which we acquire and use informative fragments, as well as category knowledge itself, have remained unclear. Here we describe the methods by which we adapted the relevant human psychophysical methods to awake, behaving monkeys and replicated key previous psychophysical results. This establishes awake, behaving monkeys as a useful system for future neurophysiological studies not only of informative fragments in particular, but also of object categorization and category learning in general.

Show MeSH