This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation

no code implementations • 4 Feb 2023 • Dimitrios E. Diamantis, Panagiota Gatoula, Anastasios Koulaouzidis, Dimitris K. Iakovidis

Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems.

Data Augmentation Image Generation

E Pluribus Unum Interpretable Convolutional Neural Networks

no code implementations • Scientific Reports 2023 • George Dimas, Eirini Cholopoulou, Dimitris K. Iakovidis

The adoption of Convolutional Neural Network (CNN) models in high-stake domains is hindered by their inability to meet society's demand for transparency in decision-making.

Decision Making

Fuzzy Pooling

no code implementations • 12 Feb 2022 • Dimitrios E. Diamantis, Dimitris K. Iakovidis

Convolutional Neural Networks (CNNs) are artificial learning systems typically based on two operations: convolution, which implements feature extraction through filtering, and pooling, which implements dimensionality reduction.

Classification Dimensionality Reduction +1

SalSum: Saliency-based Video Summarization using Generative Adversarial Networks

no code implementations • 20 Nov 2020 • George Pantazis, George Dimas, Dimitris K. Iakovidis

The proposed method is evaluated in comparison to state-of-the-art VS approaches on the benchmark dataset VSUMM.

Generative Adversarial Network Video Summarization