Filters








4 Hits in 0.58 sec

Symbolic image detection using scene and knowledge graphs [article]

Nasrin Kalanat, Adriana Kovashka
2022 arXiv   pre-print
Sometimes the meaning conveyed by images goes beyond the list of objects they contain; instead, images may express a powerful message to affect the viewers' minds. Inferring this message requires reasoning about the relationships between the objects, and general common-sense knowledge about the components. In this paper, we use a scene graph, a graph representation of an image, to capture visual components. In addition, we generate a knowledge graph using facts extracted from ConceptNet to
more » ... n about objects and attributes. To detect the symbols, we propose a neural network framework named SKG-Sym. The framework first generates the representations of the scene graph of the image and its knowledge graph using Graph Convolution Network. The framework then fuses the representations and uses an MLP to classify them. We extend the network further to use an attention mechanism which learn the importance of the graph representations. We evaluate our methods on a dataset of advertisements, and compare it with baseline symbolism classification methods (ResNet and VGG). Results show that our methods outperform ResNet in terms of F-score and the attention-based mechanism is competitive with VGG while it has much lower model complexity.
arXiv:2206.04863v1 fatcat:g6z6fkjz4jaahplt2fqrsk4l6y

Extracting Actionable Knowledge from Social Networks using Structural Features

Nasrin Kalanat, Eynollah Khanjari, Alireza Khanshan
2020 IEEE Access  
NASRIN KALANAT received the M.Sc. degree in computer engineering from the Isfahan University of Technology.  ...  Afterward, in [7] , [8] Kalanat et al. proposed a method to find profitable action rules.  ... 
doi:10.1109/access.2020.2983146 fatcat:pzmygpvgbrcedi3utq72csb6vm

Data Mining Methods for Rule Designing and Rule Triggering in Active Database Systems

Nasrin Kalanat, Mohammad Reza Kangavari
2015 International Journal of Database Theory and Application  
Active database system has been introduced to extend the database functionality. It is capable of detecting specific events and automatically reacting to them by executing certain actions either inside or outside the database. This behavior is usually specified through Event Condition Action (ECA) rules. Rule design plays a critical role in implementing an active database system. It is not always straightforward due to lack of methodology. In this paper a new approach is presented to identify
more » ... A rules that could strongly express application semantics. Additionally an ensemble method is proposed to trigger more appropriate ECA rules when the interesting events occurred.
doi:10.14257/ijdta.2015.8.1.05 fatcat:g5dhib422jeyfboh74mkwkil5m

An Efficient Probability Estimation Decision Tree Postprocessing Method for Mining Optimal Profitable Knowledge for Enterprises with Multi-Class Customers

JanapatyI Naga Muneiah, Ch D V SubbaRao
2019 Inteligencia Artificial  
Instead of class probabilities, Nasrin Kalanat et al. introduced [10, 11] fuzzy based methods which makes use of fuzzy membership to mine the profitable actions from the data.  ...  Nasrin Kalanat and Eynollah Khanjari proposed a new cost-sensitive method [28] for mining actionable knowledge from graph data belonging to social networks where there can be relationships between the  ... 
doi:10.4114/intartif.vol22iss64pp63-84 fatcat:wx6fn7q5efgdhor6z3nri63v3i