Savvas Karatsiolis - Academia.edu (original) (raw)

Papers by Savvas Karatsiolis

Global Earth Observation (EO) is becoming increasingly important in understanding and addressing ... more Global Earth Observation (EO) is becoming increasingly important in understanding and addressing critical aspects of life on our planet about environmental issues, natural disasters, sustainable development and others. EO plays a key role in making informed decisions on applying or reforming land use, responding to disasters, shaping climate adaptation policies etc. EO is also becoming a useful tool for helping professionals make the most profitable decisions, e.g., in real estate or the investment sector. Finding similarities in landscapes may provide useful information regarding applying contiguous policies, taking alike decisions or learning from best practices on events and happenings that have already occurred in similar landscapes in the past. However, current applications of similar landscape retrieval are limited by moderate performance and the need for time-consuming and costly annotations. We propose splitting the similar landscape retrieval task into a set of smaller task...

Cornell University - arXiv, Sep 19, 2022

The widespread use of black-box AI models has raised the need for algorithms and methods that exp... more The widespread use of black-box AI models has raised the need for algorithms and methods that explain the decisions made by these models. In recent years, the AI research community is increasingly interested in models' explainability since black-box models take over more and more complicated and challenging tasks. Explainability becomes critical considering the dominance of deep learning techniques for a wide range of applications, including but not limited to computer vision. In the direction of understanding the inference process of deep learning models, many methods that provide human comprehensible evidence for the decisions of AI models have been developed, with the vast majority relying their operation on having access to the internal architecture and parameters of these models (e.g., the weights of neural networks). We propose a model-agnostic method for generating saliency maps that has access only to the output of the model and does not require additional information such as gradients. We use Differential Evolution (DE) to identify which image pixels are the most influential in a model's decision-making process and produce class activation maps (CAMs) whose quality is comparable to the quality of CAMs created with model-specific algorithms. DE-CAM achieves good performance without requiring access to the internal details of the model's architecture at the cost of more computational complexity.

2022 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)

Artificial Neural Networks and Machine Learning – ICANN 2018, 2018

As a new classification platform, deep learning has recently received increasing attention from r... more As a new classification platform, deep learning has recently received increasing attention from researchers and has been successfully applied to many domains. In some domains, like bioinformatics and robotics, it is very difficult to construct a large-scale well-annotated dataset due to the expense of data acquisition and costly annotation, which limits its development. Transfer learning relaxes the hypothesis that the training data must be independent and identically distributed (i.i.d.) with the test data, which motivates us to use transfer learning to solve the problem of insufficient training data. This survey focuses on reviewing the current researches of transfer learning by using deep neural network and its applications. We defined deep transfer learning, category and review the recent research works based on the techniques used in deep transfer learning.

IEEE Transactions on Geoscience and Remote Sensing

Despite the plethora of successful super-resolution (SR) reconstruction (SRR) models applied to n... more Despite the plethora of successful super-resolution (SR) reconstruction (SRR) models applied to natural images, their application to remote sensing imagery tends to produce poor results. Remote sensing imagery is often more complicated than natural images, has its peculiarities such as being of lower resolution, contains noise, and often depicts large textured surfaces. As a result, applying nonspecialized SRR models like the enhanced SR generative adversarial network (ESRGAN) on remote sensing imagery results in artifacts and poor reconstructions. To address these problems, we propose a novel strategy for enabling an SRR model to output realistic remote sensing images: instead of relying on feature-space similarities as a perceptual loss, the model considers pixel-level information inferred from the normalized digital surface model (nDSM) of the image. This allows the application of better-informed updates during the training of the model which sources from a task (elevation map inference) that is closely related to remote sensing. Nonetheless, the nDSM auxiliary information is not required during production, i.e., the model infers an SR image without additional data. We assess our model on two remotely sensed datasets of different spatial resolutions that also contain the DSMs of the images: the Data Fusion 2018 Contest (DFC2018) dataset and the dataset containing the national LiDAR flyby of Luxembourg. We compare our model with ESRGAN, and we show that it achieves better performance and does not introduce any artifacts in the results. In particular, the results for the high-resolution DFC2018 dataset are realistic and almost indistinguishable from the ground-truth images.

A challenge of the computer vision community is to understand the semantics of an image that will... more A challenge of the computer vision community is to understand the semantics of an image that will allow for higher quality image generation based on existing high-level features and better analysis of (semi-) labeled datasets. Categorical labels aggregate a huge amount of information into a binary value which conceals valuable high-level concepts from the Machine Learning models. Towards addressing this challenge, this paper introduces a method, called Occlusion-based Latent Representations (OLR), for converting image labels to meaningful representations that capture a significant amount of data semantics. Besides being informationrich, these representations compose a disentangled low-dimensional latent space where each image label is encoded into a separate vector. We evaluate the quality of these representations in a series of experiments whose results suggest that the proposed model can capture data concepts and discover data interrelations.

Animal Biotelemetry, 2021

Background The ability to automatically count animals is important to design appropriate environm... more Background The ability to automatically count animals is important to design appropriate environmental policies and to monitor their populations in relation to biodiversity and maintain balance among species. Out of all living mammals on Earth, 60% are livestock, 36% humans, and only 4% are animals that live in the wild. In a relatively short period, development of human civilization caused a loss of 83% of wildlife and 50% of plants. The rate of species extinction is accelerating. Traditional wildlife surveys provide rough population estimates. However, emerging technologies, such as aerial photography, allow to perform large-scale surveys in a short period of time with high accuracy. In this paper, we propose the use of computer vision, through deep learning (DL) architecture, together with aerial photography and density maps, to count the population of Steller sea lions and African elephants with high precision. Results We have trained two deep learning models, a basic UNet witho...

ArXiv, 2020

A challenge of the computer vision community is to understand the semantics of an image, in order... more A challenge of the computer vision community is to understand the semantics of an image, in order to allow image reconstruction based on existing high-level features or to better analyze (semi-)labelled datasets. Towards addressing this challenge, this paper introduces a method, called Occlusion-based Latent Representations (OLR), for converting image labels to meaningful representations that capture a significant amount of data semantics. Besides being informational rich, these representations compose a disentangled low-dimensional latent space where each image label is encoded into a separate vector. We evaluate the quality of these representations in a series of experiments whose results suggest that the proposed model can capture data concepts and discover data interrelations.

Estimating the height of buildings and vegetation in single aerial images is a challenging proble... more Estimating the height of buildings and vegetation in single aerial images is a challenging problem. A task-focused Deep Learning (DL) model that combines architectural features from successful DL models (U-NET and Residual Networks) and learns the mapping from a single aerial imagery to a normalized Digital Surface Model (nDSM) was proposed. The model was trained on aerial images whose corresponding DSM and Digital Terrain Models (DTM) were available and was then used to infer the nDSM of images with no elevation information. The model was evaluated with a dataset covering a large area of Manchester, UK, as well as the 2018 IEEE GRSS Data Fusion Contest LiDAR dataset. The results suggest that the proposed DL architecture is suitable for the task and surpasses other state-of-the-art DL approaches by a large margin.

Estimating the heightmaps of buildings and vegetation in single remotely sensed images is a chall... more Estimating the heightmaps of buildings and vegetation in single remotely sensed images is a challenging problem. Effective solutions to this problem can comprise the stepping stone for solving complex and demanding problems that require 3D information of aerial imagery in the remote sensing discipline, which might be expensive or not feasible to require. We propose a task-focused Deep Learning (DL) model that takes advantage of the shadow map of a remotely sensed image to calculate its heightmap. The shadow is computed efficiently and does not add significant computation complexity. The model is trained with aerial images and their Lidar measurements, achieving superior performance on the task. We validate the model with a dataset covering a large area of Manchester, UK, as well as the 2018 IEEE GRSS Data Fusion Contest Lidar dataset. Our work suggests that the proposed DL architecture and the technique of injecting shadows information into the model are valuable for improving the h...

2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)

Over the past couple of decades, the number of wildfires and area of land burned around the world... more Over the past couple of decades, the number of wildfires and area of land burned around the world has been steadily increasing, partly due to climatic changes and global warming. Therefore, there is a high probability that more people will be exposed to and endangered by forest fires. Hence there is an urgent need to design pervasive systems that effectively assist people and guide them to safety during wildfires. This paper presents EscapeWildFire, a mobile application connected to a backend system which models and predicts wildfire geographical progression, assisting citizens to escape wildfires in real-time. A small pilot indicates the correctness of the system. The code is open-source; fire authorities around the world are encouraged to adopt this approach.

Computer Analysis of Images and Patterns

XIV Mediterranean Conference on Medical and Biological Engineering and Computing 2016, 2016

An original classification algorithm is proposed for dealing with extremely imbalanced datasets t... more An original classification algorithm is proposed for dealing with extremely imbalanced datasets that often appear in biomedical problems. Its originality comes from the way a neural network is trained in order to get a decent hypothesis out of a dataset that comprises of a huge sized majority class and a tiny size minority class. This situation is especially probable when forming machine learning databases describing rare medical conditions. The algorithm is tested on a large dataset in order to predict the risk of preeclampsia in pregnant women. Conventional machine learning algorithms tend to provide poor hypothesis for extremely imbalanced datasets by favoring the majority class. The proposed algorithm is not trained on the basis of the mean squared error objective function and thus avoids the overwhelming effect of the highly asymmetric class sizes. The methodology provides preeclampsia detection rate of 49% and normal case detection rate slightly above 76%.

Communications in Computer and Information Science, 2019

This paper describes preliminary work in the recent promising approach of generating synthetic tr... more This paper describes preliminary work in the recent promising approach of generating synthetic training data for facilitating the learning procedure of deep learning (DL) models, with a focus on aerial photos produced by unmanned aerial vehicles (UAV). The general concept and methodology are described, and preliminary results are presented, based on a classification problem of fire identification in forests as well as a counting problem of estimating number of houses in urban areas. The proposed technique constitutes a new possibility for the DL community, especially related to UAV-based imagery analysis, with much potential, promising results, and unexplored ground for further research.

Natural forests are complex ecosystems whose tree species distribution and their ecosystem functi... more Natural forests are complex ecosystems whose tree species distribution and their ecosystem functions are still not well understood. Sustainable management of these forests is of high importance because of their significant role in climate regulation, biodiversity, soil erosion and disaster prevention among many other ecosystem services they provide. In Japan particularly, natural forests are mainly located in steep mountains, hence the use of aerial imagery in combination with computer vision are important modern tools that can be applied to forest research. Thus, this study constitutes a preliminary research in this field, aiming at classifying tree species in Japanese mixed forests using UAV images and deep learning in two different mixed forest types: a black pine (Pinus thunbergii)-black locust (Robinia pseudoacacia) and a larch (Larix kaempferi)-oak (Quercus mongolica) mixed forest. Our results indicate that it is possible to identify black locust trees with 62.6 % True Positiv...

IEEE Transactions on Neural Networks and Learning Systems

Neural Computing and Applications