A 3D-CAE-CNN model for Deep Representation Learning of 3D images. (August 2022)

• Record Type: Journal Article
• Authors: Pintelas, Emmanuel; Pintelas, Panagiotis
• Is Part Of: Engineering applications of artificial intelligence. Volume 113(2022)
• Publication Date: 2022-08
• Subjects: Deep Representation Learning -- Convolutional AutoEncoders -- Deep convolutional neural networks -- 3D image classification -- DeepFake video detection -- 3D Pneumonia detection; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2022.104978
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Deep Representation Learning technologies based on supervised Convolutional Neural Networks (CNNs) have attained significant interest mainly due to their superior performance for learning abstract and robust features used in object detection and image classification tasks. However, to efficiently train such models requires a large number of labeled instances especially when these instances are high dimensional such as for 3-Dimensional (3D) Image inputs. Due to this extra dimension the dimensionality of such instances increases drastically. Therefore, the utilization of Unsupervised CNNs topologies such 3D Convolutional AutoEncoders (3D-CAE) have also been proposed. CAEs can learn features (and later used for classification tasks using common machine learning classifiers), without relying on instance labels and thus they are not prone to label limitation. Nevertheless, it is not clear if the features that CAEs learn, are relevant regarding the classification or object detection task since these features are learned via no target output class. For these reasons, in this work we combine 3D-CAE and 3D-CNN to work synergistically together in order to build a hybrid deep representation learning framework model which exploits the advantages of both unsupervised and supervised representation/feature learning approaches, applied on 3D Image inputs. In order to evaluate our strategy, we performed extensive experimental simulations for the DeepFake and Pneumonia detection problems utilizing Video and 3D Scans datasets respectively. Our proposed framework outperformed all the other utilized frameworks, revealing the efficiency of our applied methodology. Highlights: We propose the idea of combining 3D-CAE-UFL and 3D-CNN-SFL approaches in order to create efficient and high quality deep learning representations for 3D images. We provide a scientific qualitative explanation highlighting the reasons and the advantages of combining these two approaches, in order to justify the efficiency of the proposed methodology. We propose the idea of combining both LReLU and ReLU activation functions into the same network, in such a way so as to exploit the advantages of both functions. We propose a novel 3D image classification framework which exhibits a high performance applied on DeepFake and Pneumonia detection, outperforming other state of the art approaches. We conduct extensive experimental simulations providing a comprehensive comparison study for 3D image classification application case study scenarios.