Search results for "image"

Zero-shot Semantic Segmentation using Relation Network

2021

Zero-shot learning (ZSL) is widely studied in recent years to solve the problem of lacking annotations. Currently, most studies on ZSL are for image classification and object detection. But, zero-shot semantic segmentation, pixel level classification, is still at its early stage. Therefore, this work proposes to extend a zero-shot image classification model, Relation Network (RN), to semantic segmentation tasks. We modified the structure of RN based on other state-of-the-arts semantic segmentation models (i.e. U-Net and DeepLab) and utilizes word embeddings from Caltech-UCSD Birds 200-2011 attributes and natural language processing models (i.e. word2vec and fastText). Because meta-learning …

Latency-Oblivious Distributed Task Scheduling for Mobile Edge Computing

2018

Mobile Edge Computing (MEC) is emerging as one of the effective platforms for offloading the resource- and latency-constrained computational services of modern mobile applications. For latency- and resource-constrained mobile devices, the important issues include: 1) minimize end-to-end service latency; 2) minimize service completion time; 3) high quality-of-service (QoS) requirement to offload the complex computational services. To address the above issues, a latencyoblivious distributed task scheduling scheme is designed in this work to maximize the QoS performance and goodput for the MEC services. Unlike most of the existing works, we consider the latency-oblivious property of different …

Modeling and Mitigating Errors in Belief Propagation for Distributed Detection

2021

We study the behavior of the belief-propagation (BP) algorithm affected by erroneous data exchange in a wireless sensor network (WSN). The WSN conducts a distributed multidimensional hypothesis test over binary random variables. The joint statistical behavior of the sensor observations is modeled by a Markov random field whose parameters are used to build the BP messages exchanged between the sensing nodes. Through linearization of the BP message-update rule, we analyze the behavior of the resulting erroneous decision variables and derive closed-form relationships that describe the impact of stochastic errors on the performance of the BP algorithm. We then develop a decentralized distribute…

Combining Social Service and Healthcare as the First Country in the World : Exploring the Impacts on Information Systems

2018

The Finnish government has decided to implement a reform in the social and healthcare system by combining the two in the future. There are several drivers for this change that have been identified. Large number of information systems that are not interoperable, challenges in data management and isolated service offering are the most significant ones. In the government’s strategy for the future the digitalisation will provide tools to solve these challenges. In this paper the landscape is outlined and the architecture choices are discussed. The enterprise architecture is applied for the largest county comprising of over 1900 social service and healthcare related information systems. The targ…

Kaksi maailmaa? : sukupuoli Mika Waltarin kirjailijakuvassa, teksteissä ja niiden vastaanotossa 1925-1939

2013

”Siellä missä miehiä tehdään” – Mika Waltarin nuoruuden tuotanto ja sen vastaanotto sukupuolten rakentamisen paikkoina

2013

Arvio teoksesta Järvelä, Juha: Waltari ja sukupuolten maailmat. Helsinki: Avain. 2013. 265 s.

Kaunokirjallisuus ja siitä käydyt keskustelut sukupuolihistorian lähteenä: Lectio praecursoria 13.6.2013

2013

Juha Järvelän Suomen historian väitöskirja Kaksi maailmaa? Sukupuoli Mika Waltarin kirjailijakuvassa, teksteissä ja niiden vastaanotossa 1925–1939 tarkastettiin 13.6.2013 Jyväskylän yliopistossa. Vastaväittäjänä toimi dosentti Maarit Leskelä-Kärki Turun yliopistosta ja kustoksena professori Kustaa H.J. Vilkuna Jyväskylän yliopistosta.

Faktaa fiktion keinoin : nonfiktiota Helsingin Sanomien Kuukausiliitteessä ja Imagessa

2001

Analyse multirésolution pour la recherche et l'indexation d'images par le contenu dans les bases de données images - Application à la base d'images p…

2005

Recent content-based image retrieval systems offer an interactive visual browsing of images databases. These methods perform a classification of images (offline) into a search tree for users browsing (online). This approach shows three main problems:1) The size of decriptor vector (n>100) makes distance computing sensitive to dimensionality curse,2) Having many different kinds of attributes into descriptor vector does not help classification,3) In general, classification does not take in consideration users' search context. In this work, we propose a method based on building hierarchical signatures having small increasing sizes, this allows to take users' search context into consideration. …

Monitoring the evolution of coastline in Sicily by aerial photos and high resolution satellite imagery

2009

Italy is one of the country more exposed to coasts erosion: about 1200 km of coasts on about 8350 km are moved back of about 25 m in the last 50 years. Particularly Sicily have about 300 km of coasts moved back. There are various reasons of this phenomenon: natural factor (climatic or physical-oceanographic) with environmental and economic effects. The paper reviews the activity of the Dipartimento di Rappresentazione of University of Palermo in cooperation with the Regional Agency of Environmental Protection (ARPA) about the coastline evolution in Sicily. The research is carried out on selected physiographic units marked out by important erosive phenomena. The coastline trend are reconstru…