Bottom-Up Attention Guidance For Recurrent Image Recognition
Keywords
Deep neural networks; Gaze; Image recognition; Recurrent neural networks; Saliency
Abstract
This paper presents a recurrent neural network architecture, guided by the bottom-up attention, for the recognition task. The proposed architecture processes an input image as a sequence of selectively chosen patches. The patches are chosen from the salient regions of the input image. Using human driven saliency maps from gaze, the benefit of such a selection process is first shown. Next, the performance of computational models of bottom-up attention are assessed as alternative to human attention.
Publication Date
8-29-2018
Publication Title
Proceedings - International Conference on Image Processing, ICIP
Number of Pages
3004-3008
Document Type
Article; Proceedings Paper
Personal Identifier
scopus
DOI Link
https://doi.org/10.1109/ICIP.2018.8451537
Copyright Status
Unknown
Socpus ID
85062916136 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/85062916136
STARS Citation
Tavakoli, Hamed R.; Borji, Ali; Anwer, Rao Muhammad; Rahtu, Esa; and Kannala, Juho, "Bottom-Up Attention Guidance For Recurrent Image Recognition" (2018). Scopus Export 2015-2019. 9564.
https://stars.library.ucf.edu/scopus2015/9564