Comparative study of visual saliency maps in the problem of classification of architectural images with Deep CNNs

Obeso, Abraham Montoya; Benois-Pineau, Jenny; Guissous, Kamel; Gouet-Brunet, Valérie; Vázquez, Mireya Saraí García; Acosta, Alejandro

doi:10.1109/ipta.2018.8608125

Cited by 5 publications

(2 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence in this paper, we compare the presented white-box methods on the datasets with human gaze fixations recorded in a task-driven visual experiment of recognition of architectural styles of Mexican architectural heritage, introduced in Ref. 56 From this table, one can see that FEM is the most similar to the GFDM in SIM-L1 norm and that Grad-CAM has the highest correlation with the GFDM. The methods belong to the linearization category (Sec.…”

Section: Human Perception-based Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Review of white box methods for explanations of convolutional neural networks in image classification tasks

2021

Self Cite

View full text Add to dashboard Cite

In recent years, deep learning has become prevalent to solve applications from multiple domains. Convolutional neural networks (CNNs) particularly have demonstrated state-ofthe-art performance for the task of image classification. However, the decisions made by these networks are not transparent and cannot be directly interpreted by a human. Several approaches have been proposed to explain the reasoning behind a prediction made by a network. We propose a topology of grouping these methods based on their assumptions and implementations. We focus primarily on white box methods that leverage the information of the internal architecture of a network to explain its decision. Given the task of image classification and a trained CNN, our work aims to provide a comprehensive and detailed overview of a set of methods that can be used to create explanation maps for a particular image, which assign an importance score to each pixel of the image based on its contribution to the decision of the network. We also propose a further classification of the white box methods based on their implementations to enable better comparisons and help researchers find methods best suited for different scenarios.

show abstract

Section: Human Perception-based Evaluationmentioning

confidence: 99%

“…Comparison of explanation maps of different methods with the GFDM for the VGG16 network trained on the MexCulture dataset 56. Pineau and Zemmari: Review of white box methods for explanations of convolutional neural.…”

mentioning

confidence: 99%