Sök:

Sökresultat:

14085 Uppsatser om Machine learning based image classification of electronic components - Sida 1 av 939

Perceptuell identifiering av musikgenre

This is a study in the field of Music Information Retrieval (MIR) concerning classification of musicgenre. The study uses a music listening experiment and machine learning methods to determinewhether or not the drum sounds from different songs contain sufficient information to classifygenre.MIR methods are used to extract data from the audio signals. The data is used together with theclassification results of the listening experiment to train different machine learning algorithms toclassify the songs in different genres.The results indicate that using only drum sounds for genre classifications works well, both forlisteners and for machine learning methods..

Påverkan på elektronik och sensorer i radiologisk miljö

The objective of this thesis is to make a training material for KSU, which shows how radiation affects the different electronic components in a radiological environment.All electronic systems in the presence of high ionizing radiation can experience difficulties and be destroyed.This influence may compromise the control system and thus the safety of a nuclear reactor plant.To achieve this objective requires a description of electronic components and sensors, to better understand its function.This thesis will also show the possible solutions available that the author found interesting.Organizations such as NASA, CERN and ESA have worked to develop electronic components immune to radiation. Those organizations have also developed a test protocol that guarantees to a certain extent the security of electronic components in a radiological environment.Since we know how the material reacts with the radiation, it could develop a material that can be resistant to radiation or like CMOS technology has proposed, to reduce the component to avoid the storage of charge in the oxide.Another solution can be the GRAPHENE. Its features will revolutionize the electronics and provide answer to the radiation problems, Graphene will replace Silicium and the electronic components will only have a few nanometers dimensions so the oxide area also reduces, and in this way avoid the known problem of storage of charge..

Kinesisk klassifikation: daoistiska, konfucianska och maoistiska influenser i fem kinesiska klassifikationssystem

The aim of this thesis is to mediate to the reader an image of Chinese classification systems and in what way the methods of classification have progressed during time. The tradition of classification in China is very old and dates back more than two thousand years, but it wasnt until the downfall of the dynastic era in the beginning of the 20th century, that Chinese libraries and methods of classification were modernized. This thesis also identifies the development of Chinese librarianship which is of old age since Chinese literature stretches three thousand years back. The main question is in what way the values of three different philosophies are reflected in five classification systems compiled between 1953-1975. The three philosophies are Daoism, Confucianism and Maoism/Mao Zedong Thought.

Maskininlärningsmetoder för bildklassificering av elektroniska komponenter

Micronic Mydata AB utvecklar och tillverkar maskiner för att automatisk montera elektroniska komponenter på kretskort, s.k. ?Pick and place? (PnP) maskiner. Komponenterna blir lokaliserade och inspekterade optiskt innan de monteras på kretskorten, för att säkerhetsställa att de monteras korrekt och inte är skadade. En komponent kan t.ex.

Consumer Perceptions of Store Image: A study of Ikea and Ilva in Sweden and Denmark

Purpose: The aim of this study is to inquire into the congruency of store image perceptions, projected by Ikea and Ilva, and how Danish and Swedish consumers perceive the aforementioned stores? images in their respective domestic market. Also, the aim sets out to evaluate the success of these stores? images on consumers and on transference to host markets. Further, the aim is to discern specific store image attributes captivating Swedish and Danish consumers.

Säkerhetsklassificering av IT-system på Fläkt Woods AB

This report presents a method for modelling a computer system from a security perspective. The questions that are going to be treated are:? What defines a secure system and how does the company relate to these factors?? What are the threats today based on hardware/software, human factors and company routines/policies?? What measures should be taken for the organisation to reach a higher level of security for their systems?? How do we develop a method for classification of security and what components should it contain?? What changes are reasonable and necessary with the respect to the company?s resources?The report has been done through interviews and analysis of existing systems on Fla?kt Woods AB. From analysis of material, the aspects judged relevant to the subject and to the company?s needs, have been compiled to a document.

Konstruktion av en Telesvarsterminal

Telia Telesvar is an automatic anwering machine that is provided in your stationary telephone at home. The drawback of this is that you cannot see if anyone has called you. This report describes how a terminal for Telias Telesvar could be designed. Telesvar terminal will show in a display if anyone called you and show the incoming number. It will also play the spoken message in a built in speaker.

Konstruktion av lapputmatningsmaskin

The vision of this thesis was to create a concept that can be used as a base for amachine construction. The purpose of this project was to use different methods todevelop a concept that on the best way meets the sat up requirements. This projecthas been divided into three stages/parts; principle, primary and the final solutionstage.This thesis has been conducted in collaboration with Västkustägg AB. It is aconstruction project designed to construct a machine that will automatically putlabels on the egg packages. Västkust AB wanted to automate this process.

Klassifikationssystem på webben: Traditionella och nya

The aim of this thesis is to investigate the opportunities of classification of web pages. Two classification systems are compared; a library classification system traditional and a system that is developed for use on the web new. UDC is chosen to represent the traditional systems, and Yahoo! represents a so called new system. UDC is used by NISS Directory of networked resources, which is a directory that uses the system for organisation of electronic resources. Both systems are analysed from two different perspectives, a multidisciplinary and a timeperspective.

Klassifikationens roll på www: en studie av webbmiljöns påverkan på klassifikationens principer och funktioner

Principles that guide how a classification system should be structured have existed for more than a century. Traditional classification principles can be of great value for the design of classification schemes directories on the web. However, there are differences between the library context in which these principles originated, and the web. This thesis aims at investigating how changes in the context of classification affect the application of traditional principles of classification on the web. Another purpose is to see what implications this has for the design of classification schemes on the web.

Bild- och formskapande i förskola och förskoleklass

The aim of our graduate work is to study the importance of image and shape?s in children's development and learning and also the preschool teacher?s views on image and shapes. Our research aims to explore the value of image and shapes creation and how it is expressed among children and in classroom activities. We are also interested in finding out how it differs from adults' views and what materials teachers use to promote the image and shapes creation in the curriculum. In our own experience, image and shapes activities are often structured differently in different curriculum activities.

Kontextbaserad utveckling av klassifikationssystem: Värdet av ett användarperspektiv i en beståndsorganiserande process på specialbibliotek

This thesis treats a new perspective in the theory of classification, the context-based approach. The aim is to find out what kind of value this new approach can have in the process of creating a classification scheme for a special library collection. The first question to be answered is how the context-based approach works in a real life. The second question investigates which role the research in this particular area might have. Illustrating a real-life situation, the unit of Rättsmedicinalverket in Linköping serves as the object of study.

Testeffekten som medel för att förbättra läsförståelse?

Swedish school children exhibit deteriorating knowledge results in terms of reading comprehension. Previous research shows that a method to increase the level of knowledge is the use of evidence-based learning techniques. Test based learning generates what is commonly referred to as a test effect, which proved to strengthen the relationship between learning and memory, as well as being applicable to various test formats. The purpose of this study was to investigate whether test based learning, in terms of learning english words leads to increased performance in a reading comprehension test, more specifically a cloze test. The study was based on a within-group design and was conducted three times over four weeks.

Att klassificera populärmusik: en studie med förankring i SAB-systemet

The aim of this paper is to examine how The Swedish Library Classification System, also known as the ?SAB-system?, has classified popular music. The investigation is limited to the classification category of ?Yx Jazz, Rock and Popular Music? within the SAB-system. The development of this category is studied.

Automatisk genreklassifikation : en experimentell studie

This thesis aims at examining to what extent a few, algorithmically very easily extractable document features can be used to classify electronic documents according to genre. A set of experiments is therefore carried out, using only 11 such simple features in an attempt to classify 84 documents belonging to electronic academic journals into three manually identified genres: table of contents, article, and review. The 11 features are also divided into three sets, containing metrics of words and sentences; punctuation marks; and URL links, respectively. The performance when using these sets of features is then measured with regard to classification accuracy, using a k-NN classifier, four different values of k (1, 3, 5, 7), and both leave-one-out and 10-fold cross-validation. Best results are achieved when using all three feature sets (i.e.

1 Nästa sida ->