A survey on deep learning in image polarity detection : balancing generalization performances and computational costs

• Main Authors: Ragusa, Edoardo, Cambria, Erik, Zunino, Rodolfo, Gastaldo, Paolo
• Other Authors: School of Computer Science and Engineering
• Format: Article
• Language: English
• Published: 2020
• Subjects: Engineering::Computer science and engineering; Convolutional Neural Networks; Deep Learning
• Online Access: https://hdl.handle.net/10356/142828
• Institution: Nanyang Technological University

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...