A survey on deep learning in image polarity detection : balancing generalization performances and computational costs
Deep convolutional neural networks (CNNs) provide an effective tool to extract complex information from images. In the area of image polarity detection, CNNs are customarily utilized in combination with transfer learning techniques to tackle a major problem: The unavailability of large sets of label...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/142828 |
| _version_ | 1826117391101198336 |
|---|---|
| author | Ragusa, Edoardo Cambria, Erik Zunino, Rodolfo Gastaldo, Paolo |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Ragusa, Edoardo Cambria, Erik Zunino, Rodolfo Gastaldo, Paolo |
| author_sort | Ragusa, Edoardo |
| collection | NTU |
| description | Deep convolutional neural networks (CNNs) provide an effective tool to extract complex information from images. In the area of image polarity detection, CNNs are customarily utilized in combination with transfer learning techniques to tackle a major problem: The unavailability of large sets of labeled data. Thus, polarity predictors in general exploit a pre-trained CNN as the feature extractor that in turn feeds a classification unit. While the latter unit is trained from scratch, the pre-trained CNN is subject to fine-tuning. As a result, the specific CNN architecture employed as the feature extractor strongly affects the overall performance of the model. This paper analyses state-of-the-art literature on image polarity detection and identifies the most reliable CNN architectures. Moreover, the paper provides an experimental protocol that should allow assessing the role played by the baseline architecture in the polarity detection task. Performance is evaluated in terms of both generalization abilities and computational complexity. The latter attribute becomes critical as polarity predictors, in the era of social networks, might need to be updated within hours or even minutes. In this regard, the paper gives practical hints on the advantages and disadvantages of the examined architectures both in terms of generalization and computational cost. |
| first_indexed | 2024-10-01T04:26:49Z |
| format | Journal Article |
| id | ntu-10356/142828 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T04:26:49Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | ntu-10356/1428282020-07-03T03:38:15Z A survey on deep learning in image polarity detection : balancing generalization performances and computational costs Ragusa, Edoardo Cambria, Erik Zunino, Rodolfo Gastaldo, Paolo School of Computer Science and Engineering Engineering::Computer science and engineering Convolutional Neural Networks Deep Learning Deep convolutional neural networks (CNNs) provide an effective tool to extract complex information from images. In the area of image polarity detection, CNNs are customarily utilized in combination with transfer learning techniques to tackle a major problem: The unavailability of large sets of labeled data. Thus, polarity predictors in general exploit a pre-trained CNN as the feature extractor that in turn feeds a classification unit. While the latter unit is trained from scratch, the pre-trained CNN is subject to fine-tuning. As a result, the specific CNN architecture employed as the feature extractor strongly affects the overall performance of the model. This paper analyses state-of-the-art literature on image polarity detection and identifies the most reliable CNN architectures. Moreover, the paper provides an experimental protocol that should allow assessing the role played by the baseline architecture in the polarity detection task. Performance is evaluated in terms of both generalization abilities and computational complexity. The latter attribute becomes critical as polarity predictors, in the era of social networks, might need to be updated within hours or even minutes. In this regard, the paper gives practical hints on the advantages and disadvantages of the examined architectures both in terms of generalization and computational cost. Published version 2020-07-03T03:38:15Z 2020-07-03T03:38:15Z 2019 Journal Article Ragusa, E., Cambria, E., Zunino, R., & Gastaldo, P. (2019). A survey on deep learning in image polarity detection : balancing generalization performances and computational costs. Electronics, 8(7), 783-. doi:10.3390/electronics8070783 2079-9292 https://hdl.handle.net/10356/142828 10.3390/electronics8070783 2-s2.0-85071026354 7 8 en Electronics © 2019 The Authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). application/pdf |
| spellingShingle | Engineering::Computer science and engineering Convolutional Neural Networks Deep Learning Ragusa, Edoardo Cambria, Erik Zunino, Rodolfo Gastaldo, Paolo A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title | A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title_full | A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title_fullStr | A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title_full_unstemmed | A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title_short | A survey on deep learning in image polarity detection : balancing generalization performances and computational costs |
| title_sort | survey on deep learning in image polarity detection balancing generalization performances and computational costs |
| topic | Engineering::Computer science and engineering Convolutional Neural Networks Deep Learning |
| url | https://hdl.handle.net/10356/142828 |
| work_keys_str_mv | AT ragusaedoardo asurveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT cambriaerik asurveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT zuninorodolfo asurveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT gastaldopaolo asurveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT ragusaedoardo surveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT cambriaerik surveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT zuninorodolfo surveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts AT gastaldopaolo surveyondeeplearninginimagepolaritydetectionbalancinggeneralizationperformancesandcomputationalcosts |