We live in a data-driven world
Currently, we are experiencing an unprecedented digital revolution. The potential of big data is growing rapidly as information pours in from digital platforms, wireless sensors, virtual-reality applications, and billions of mobile phones. As a result data science and Artifical Intelligence (AI) reshape all kind of industries. To take advantage of those rapid technological advances companies must integrate new capabilities into their operations and strategic vision and use them to make better and faster decisions.
Due to the growing amount of electronic data there is a need for methods capable of analyzing high-dimensional data. We explore the power of data science and Artificial Intelligence (AI) to leverage data efficiently.
EtaArgus is a research group in the space of Artificial Intelligence (AI) and Computer Vision (CV). We conceptualize and develop AI solutions for real-world problems, i.e. we conduct research with a focus on practical applicability.
We conduct research in the following fields
The goal of Machine Learning (ML) is to develop methods, that allow to automatically detect patterns in the given data, and use those patterns to predict new or unseen data.
Deep learning is a thriving area within ML, which has been introduced with the objective of moving ML closer to one of its original goals: Artificial Intelligence.
The field of Computer Vision (CV) deals with inverse problems that are used to extract high-level or low-level image content out of one or more images. CV basically seeks to automate tasks that the human visual system can do.
We use the best technologies available on the market, and we are constantly adding new ones.
One of the most studied problems in Computer Vision (CV) is stereo. Given two or more images, taken from a static scene, but from different viewpoints, stereo algorithms allow to estimate the geometry of the scene.
In this project we used deep regression networks to extract geometric information from multi-view stereo data.
Image segmentation is a classical Computer Vision (CV) problem. The task is to assigne each pixel a semantic class label.
In this project we used u-shape networks to
segment areal images from the ISPRS contest.
Image denoising or image restoration is a classical Computer Vision (CV) problem. The task is to remove the noise from the image in such a way that the original image is discernible.
With cutting edge denoising algorithms we can remove noise from images without destroying the main structure of the image.
Image inpainting is a fundamental problem in Computer Vision (CV) with great practical importance. Given an image with lost, deteriorated or simply unknown regions, the task of image inpainting is to convincingly fill-up those unknown image regions.
We utilize Convolutional Neural Networks (CNNs) to remove artifacts or unwanted objects from images.