Journal of Naval Sciences and Engineering

14

Authors: Furkan GÜRPINAR, Heysem KAYA, Sadaf AFSHAR, Hamdi DİBEKLİOĞLU, Albert Ali SALAH

This paper presents a framework for estimating the apparent age of a subject from their face image. To learn the age estimation model, we used a set of visual descriptors and their feature-level fusion to obtain a single feature vector for apparent age. For model learning, we used regression with Extreme Learning Machines (ELM) with Gaussian (RBF) kernels. We tested the proposed system with k-fold cross-validation on the dataset provided by ChaLearn Looking at People 2015 - Apparent Age Estimation challenge. By combining visual descriptors from multiple grids, we obtained a Mean Absolute Error (MAE) of 5.20.

19

Authors: Muhammed Miraç SÜZGÜN, Hilal ÖZDEMİR, Necati Cihan CAMGÖZ, Ahmet Alp KINDIROĞLU, Doğaç BAŞARAN, Cengiz TOGAY, Lale AKARUN

Sign language is the natural medium of communication for the Deaf community. In this study, we have developed an interactive communication interface for hospitals, HospiSign, using computer vision based sign language recognition methods. The objective of this paper is to review sign language based Human-Computer Interaction applications and to introduce HospiSign in this context. HospiSign is designed to meet deaf people at the information desk of a hospital and to assist them in their visit. The interface guides the deaf visitors to answer certain questions and express intention of their visit, in sign language, without the need of a translator. The system consists of a computer, a touch display to visualize the interface, and Microsoft Kinect v2 sensor to capture the users’ sign responses. HospiSign recognizes isolated signs in a structured activity diagram using Dynamic Time Warping based classifiers. In order to evaluate the developed interface, we performed usability tests and deduced that the system was able to assist its users in real time with high accuracy.

14

Authors: Erkan KIYAK, Oğuz Han TİMUŞ, Mehmet KARAYEL

For a substantial organization, managing the expansive inventory is a serious problem. One of the methods to solve this problem used widely is classifying the inventory according to some criteria and managing inventory according to this classisifation. In this study, ABC classification methods are researched and Ng's model, which is one of the most widely used, selected for further investigation. An illustrative example is presented to show the usability of the Ng's method.

27

Authors: Mukrimin Sevket GUNEY

It is a consensus to widespread use of renewable sources for disposal of environmental impact caused by fossil fuel consumption, and moreover to remedy of fossil fuels depletion. Wave power is a renewable kind of energy. Worldwide potential for wave power is enormous. Ocean appears to be an important source of wave energy. Various systems are developed and some new projects are implemented on this subject. Therefore, in this study has been presented the wave energy conversion systems in detail with a classification and various aspects.

15

Authors: Mahmut IGDE, Yusuf KAVURUCU, Alev MUTLU

Multi-relational concept discovery aims to find the relational rules that best describe the target concept. In this paper, we present a graph-based concept discovery method in Multi- Relational Data Mining. Concept rule discovery aims at finding the definition of a specific concept in terms of relations involving background knowledge. The proposed method is an improvement over a state-of-the-art concept discovery system that uses both ILP and conventional association rule mining techniques during concept discovery process. The proposed method generates graph structures with respect to data that is initially stored in a relational database and utilizes them to guide the concept induction process. A set of experiments is conducted on data sets that belong to different learning problems. The results show that the proposed method has promising results in comparison to state of the art methods.

18

Authors: Oğuz Han TİMUŞ, Erkan KIYAK

The purpose of this study is to optimize multilayer perceptron (MLP) classifier and find optimal ECG features to achieve better classification for automated sleep apnea detection. k-fold crossvalidation technique was employed for classification of apneaic events on the apnea database of the DREAMS project containing 12 whole-night Polysomnography (PSG) recordings previously examined by an expert. To achieve the best possible performance with MLP, the correlation feature selection method was utilized. The performance for apnea event diagnosis after optimization of the features and the classifier resulted almost 10% in accuracy, %7 in sensitivity and %13 in specificity.