A survey on deep learning in medical image analysis

Deep convolutional neural networks have performed remarkably well on many Computer Vision tasks. However, these networks are heavily reliant on big data to avoid overfitting. Overfitting refers to the phenomenon when a network learns a function with very high variance such as to perfectly model the training data. Unfortunately, many application domains do not have access to big data, such as medical image analysis. This survey focuses on Data Augmentation, a data-space solution to the problem of limited data. Data Augmentation encompasses a suite of techniques that enhance the size and quality of training datasets such that better Deep Learning models can be built using them. The image augmentation algorithms discussed in this survey include geometric transformations, color space augmentations, kernel filters, mixing images, random erasing, feature space augmentation, adversarial training, generative adversarial networks, neural style transfer, and meta-learning. The application of augmentation methods based on GANs are heavily covered in this survey. In addition to augmentation techniques, this paper will briefly discuss other characteristics of Data Augmentation such as test-time augmentation, resolution impact, final dataset size, and curriculum learning. This survey will present existing methods for Data Augmentation, promising developments, and meta-level decisions for implementing Data Augmentation. Readers will understand how Data Augmentation can improve the performance of their models and expand limited datasets to take advantage of the capabilities of big data.

/pdf/a-survey-on-image-data-augmentation-for-deep-learning-4ru1bj2hbb.pdf

A survey on Image Data Augmentation for Deep Learning

1. Introduction and Overview. 2. Detecting Influential Observations and Outliers. 3. Detecting and Assessing Collinearity. 4. Applications and Remedies. 5. Research Issues and Directions for Extensions. Bibliography. Author Index. Subject Index.

Regression Diagnostics: Identifying Influential Data and Sources of Collinearity

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

The PASCAL Visual Object Classes Challenge

Este artigo apresenta um estudo sobre vieses gerados no aprendizado de máquina e as suas implicações na sociedade — morais, éticas e sociais. Fazemos uma releitura de um framework que posiciona os diferentes tipos de vieses nas etapas do processo de aprendizado de máquina, desde o pré-processamento, passando pela coleta dos dados, até o pós-processamento. Apresentamos um estudo de caso sobre reconhecimento facial para ilustrar os vieses que podem ser potencialmente incluídos durante estas etapas do aprendizado de máquina e as suas implicações sociais.

/pdf/vieses-no-aprendizado-de-maquina-e-suas-implicacoes-sociais-18xx6xqboh.pdf

Vieses no Aprendizado de Máquina e suas Implicações Sociais: Um Estudo de Caso no Reconhecimento Facial

Deep Learning experiments require large amounts of labeled data, but few annotated seismic datasets are available and annotation is a time-consuming, expensive activity. Synthetic modeled datasets may be a viable alternative. However, they lack the variability and intricacies of a real data signal. Moreover, methods that add colored noises are not enough to represent such variability. Thus, our goal is to produce a noise type that is characteristic of real data to create a viable synthetic dataset to train Deep Learning models. In this context, we apply the Neural StyleTransfer technique, which combines the structural content of an image with the textural style of another, to produce a synthetic data with noise characteristics extracted from a real dataset. The results show that the stylized synthetic seismic data preserves the modeled content while incorporating characteristics of some real data chosen as style, creating synthetic data with a more realistic noise profile.

/pdf/enriching-synthetic-data-with-real-noise-using-neural-style-4u5vqucor5.pdf

Enriching synthetic data with real noise using Neural Style Transfer

/pdf/automatic-detection-of-diffraction-apex-using-fully-2wrwgs316i.pdf

Automatic detection of diffraction-apex using fully convolutional networks

Natural disaster prediction is one of the main concerns of authorities globally. Disasters cause large-scale psychological, social, and economic damage; therefore, techniques to predict such events are essential to minimize their impacts. However, despite all efforts to estimate the occurrence of a disaster, making an accurate and robust forecast is a challenging task. In recent years, Deep Learning techniques have innovated several fields by learning the factors that contribute to the phenomena generation; also, biologically inspired concepts such as attention have provided cheap ways to allocate computational resources by consuming only the necessary information to solve the task. This work aims to develop and evaluate a feature-based temporal attentional model to predict the time remaining for an earthquake event by consuming seismic waves from the LANL Earthquake dataset. The proposed models comprehend four distinct architectures based on 1D-CNN, LSTM, and LSTM with self-attention mechanisms. Experimental results indicated that attention plays a vital role in learning temporal relations, with models achieving state-of-the-art results in the task.

An Attentional Model for Earthquake Prediction Using Seismic Data

Self-supervised pre-training appears as an advantageous alternative to supervised pre-trained for transfer learning. By synthesizing annotations on pretext tasks, self-supervision allows to pre-train models on large amounts of pseudo-labels before fine-tuning them on the target task. In this work, we assess self-supervision for the diagnosis of skin lesions, comparing three self-supervised pipelines to a challenging supervised baseline, on five test datasets comprising in- and out-of-distribution samples. Our results show that self-supervision is competitive both in improving accuracies and in reducing the variability of outcomes. Self-supervision proves particularly useful for low training data scenarios ($<1\,500$ and $<150$ samples), where its ability to stabilize the outcomes is essential to provide sound results.

Sandra Avila

Papers

Vieses no Aprendizado de Máquina e suas Implicações Sociais: Um Estudo de Caso no Reconhecimento Facial

Enriching synthetic data with real noise using Neural Style Transfer

Automatic detection of diffraction-apex using fully convolutional networks

An Attentional Model for Earthquake Prediction Using Seismic Data

An Evaluation of Self-Supervised Pre-Training for Skin-Lesion Analysis.