Semantic object recognition by merging decision tree with object ontology

Author(s):  
Wafa Damak ◽  
Issam Rebai ◽  
Imene Khanfir Kallel
2021 ◽  
Vol 103 (2) ◽  
Author(s):  
Haibo Sun ◽  
Feng Zhu ◽  
Yingming Hao ◽  
Shuangfei Fu ◽  
Yanzi Kong ◽  
...  

2018 ◽  
Vol 29 (7) ◽  
pp. 3023-3033 ◽  
Author(s):  
Johan N Lundström ◽  
Christina Regenbogen ◽  
Kathrin Ohla ◽  
Janina Seubert

Abstract While matched crossmodal information is known to facilitate object recognition, it is unclear how our perceptual systems encode the more gradual congruency variations that occur in our natural environment. Combining visual objects with odor mixtures to create a gradual increase in semantic object overlap, we demonstrate high behavioral acuity to linear variations of olfactory–visual overlap in a healthy adult population. This effect was paralleled by a linear increase in cortical activation at the intersection of occipital fusiform and lingual gyri, indicating linear encoding of crossmodal semantic overlap in visual object recognition networks. Effective connectivity analyses revealed that this integration of olfactory and visual information was achieved by direct information exchange between olfactory and visual areas. In addition, a parallel pathway through the superior frontal gyrus was increasingly recruited towards the most ambiguous stimuli. These findings demonstrate that cortical structures involved in object formation are inherently crossmodal and encode sensory overlap in a linear manner. The results further demonstrate that prefrontal control of these processes is likely required for ambiguous stimulus combinations, a fact of high ecological relevance that may be inappropriately captured by common task designs juxtaposing congruency and incongruency.


Author(s):  
YANG WU ◽  
NANNING ZHENG ◽  
YUANLIU LIU ◽  
ZEJIAN YUAN

This paper presents a novel research on promoting the performance and enriching the functionalities of object recognition. Instead of simply fitting various data to a few predefined semantic object categories, we propose to generate proper results for different object instances based on their actual visual appearances. The results can be fine-grained and layered categorization along with absolute or relative localization. We present a generic model based on structured prediction and an efficient online learning algorithm to solve it. Experiments on a new benchmark dataset demonstrate the effectiveness of our model and its superiority against traditional recognition methods.


Sign in / Sign up

Export Citation Format

Share Document