• Journal of radiological science and technology
• /
• 제46권4호
• /
• pp.277-285
• /
• 2023

In this study, pneumonia identification networks with the small number of layers were constructed by using chest X-ray images. The networks had similar trainable-parameters, and the performance of the trained models was quantitatively evaluated with the modification of the network architectures. A total of 6 networks were constructed: convolutional neural network (CNN), VGGNet, GoogleNet, residual network with identity blocks, ResNet with bottleneck blocks and ResNet with identity and bottleneck blocks. Trainable parameters for the 6 networks were set in a range of 273,921-294,817 by adjusting the output channels of convolution layers. The network training was implemented with binary cross entropy (BCE) loss function, sigmoid activation function, adaptive moment estimation (Adam) optimizer and 100 epochs. The performance of the trained models was evaluated in terms of training time, accuracy, precision, recall, specificity and F1-score. The results showed that the trained models with the small number of layers precisely detect pneumonia from chest X-ray images. In particular, the overall quantitative performance of the trained models based on the ResNets was above 0.9, and the performance levels were similar or superior to those based on the CNN, VGGNet and GoogleNet. Also, the residual blocks affected the performance of the trained models based on the ResNets. Therefore, in this study, we demonstrated that the object detection networks with the small number of layers are suitable for detecting pneumonia using chest X-ray images. And, the trained models based on the ResNets can be optimized by applying appropriate residual-blocks.

• Advances in concrete construction
• /
• 제14권3호
• /
• pp.185-194
• /
• 2022

For producing cement and concrete, the construction field has been encouraged by the usage of industrial soil waste (or) secondary materials since it decreases the utilization of natural resources. Simultaneously, for ensuring the quality, the analyses of the strength along with durability properties of that sort of cement and concrete are required. The prediction of strength along with other properties of High-Performance Concrete (HPC) by optimization and machine learning algorithms are focused by already available research methods. However, an error and accuracy issue are possessed. Therefore, the Enhanced Deep Neural Network (EDNN) based strength along with durability prediction of HPC was utilized by this research method. Initially, the data is gathered in the proposed work. Then, the data's pre-processing is done by the elimination of missing data along with normalization. Next, from the pre-processed data, the features are extracted. Hence, the data input to the EDNN algorithm which predicts the strength along with durability properties of the specific mixing input designs. Using the Switched Multi-Objective Jellyfish Optimization (SMOJO) algorithm, the weight value is initialized in the EDNN. The Gaussian radial function is utilized as the activation function. The proposed EDNN's performance is examined with the already available algorithms in the experimental analysis. Based on the RMSE, MAE, MAPE, and R² metrics, the performance of the proposed EDNN is compared to the existing DNN, CNN, ANN, and SVM methods. Further, according to the metrices, the proposed EDNN performs better. Moreover, the effectiveness of proposed EDNN is examined based on the accuracy, precision, recall, and F-Measure metrics. With the already-existing algorithms i.e., JO, GWO, PSO, and GA, the fitness for the proposed SMOJO algorithm is also examined. The proposed SMOJO algorithm achieves a higher fitness value than the already available algorithm.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 한국작물학회 2022년도 추계학술대회
• /
• pp.88-88
• /
• 2022

Detection of stress responses in crops is important to diagnose crop growth and evaluate yield. Also, the multi-spectral sensor is effectively known to evaluate stress caused by nutrient and moisture in crops or biological agents such as weeds or diseases. Therefore, in this experiment, multispectral images were taken by an unmanned aerial vehicle(UAV) under field condition. The experiment was conducted in the long-term fertilizer field in the National Institute of Crop Science, and experiment area was divided into different status of NPK(Control, N-deficiency, P-deficiency, K-deficiency, Non-fertilizer). Total 11 vegetation indices were created with RGB and NIR reflectance values using python. Variations in nutrient content in plants affect the amount of light reflected or absorbed for each wavelength band. Therefore, the objective of this experiment was to evaluate vegetation indices derived from multispectral reflectance data as input into machine learning algorithm for the classification of nutritional deficiency in rice. RandomForest model was used as a representative ensemble model, and parameters were adjusted through hyperparameter tuning such as RandomSearchCV. As a result, training accuracy was 0.95 and test accuracy was 0.80, and IPCA, NDRE, and EVI were included in the top three indices for feature importance. Also, precision, recall, and f1-score, which are indicators for evaluating the performance of the classification model, showed a distribution of 0.7-0.9 for each class.

• The Journal of the Korea Contents Association
• /
• 제11권12호
• /
• pp.511-521
• /
• 2011

As the use of image input and output devices increases, the importance of extracting text area in color images is also increasing. In this paper, in order to extract text area of the images efficiently, we present a text area extraction method for color images based on labeling and gradient difference method. The proposed method first eliminates non-text area using the processes of labeling and filtering. After generating the candidates of text area by using the property that is high gradient difference in text area, text area is extracted using the post-processing of noise removal and text area merging. The benefits of the proposed method are its simplicity and high accuracy that is better than the conventional methods. Experimental results show that precision, recall and inverse ratio of non-text extraction (IRNTE) of the proposed method are 99.59%, 98.65% and 82.30%, respectively.

• Proceedings of the Korea Contents Association Conference
• /
• 한국콘텐츠학회 2006년도 춘계 종합학술대회 논문집
• /
• pp.140-143
• /
• 2006

In this paper, we describe a question-answering system which extracts answers for the superlative questions which include the superlative words such as "the most", "the best", "the first", "the largest", "the least", and so on. The superlative questions are composed of four main components and others. Four main components are the superlative word, answer type, regional information, and a verb modified by the superlative word. We classify the superlative words into two types as to whether the verb has to be needed to be a question or not. The superlative word, answer type and regional information are essential elements to extract answer for all superlative questions. But the verb may be an essential element by the type of superlative word. Our system analyzes input question, and finds four main components of the superlative question. Also, our system searches relative documents and candidate sentences using them, and extracts answers from candidate sentences. Empirical result shows that our system has high precision and high recall for the superlative questions.

• Journal of the Korea Society for Simulation
• /
• 제27권2호
• /
• pp.1-10
• /
• 2018

This paper presents an autonomous detection and aiming point computation of a battle tank by using RGB images. Maximally stable extremal regions algorithm was implemented to find features of the tank, which are matched with images extracted from streaming video to figure out the region of interest where the tank is present. The median filter was applied to remove noises in the region of interest and decrease camouflage effects of the tank. For the tank segmentation, k-mean clustering was used to autonomously distinguish the tank from its background. Also, both erosion and dilation algorithms of morphology techniques were applied to extract the tank shape without noises and generate the binary image with 1 for the tank and 0 for the background. After that, Sobel's edge detection was used to measure the outline of the tank by which the aiming point at the center of the tank was calculated. For performance measurement, accuracy, precision, recall, and F-measure were analyzed by confusion matrix, resulting in 91.6%, 90.4%, 85.8%, and 88.1%, respectively.

• Journal of the Korea Society of Computer and Information
• /
• 제17권10호
• /
• pp.25-33
• /
• 2012

In this paper, we propose a rule-based model to correct errors in a speech recognition result in the mobile device environment. The proposed model considers the mobile device environment with limited resources such as processing time and memory, as follows. In order to minimize the error correction processing time, the proposed model removes some processing steps such as morphological analysis and the composition and decomposition of syllable. Also, the proposed model utilizes the longest match rule selection method to generate one error correction candidate per point, assumed that an error occurs. For the purpose of deploying memory resource, the proposed model uses neither the Eojeol dictionary nor the morphological analyzer, and stores a combined rule list without any classification. Considering the modification and maintenance of the proposed model, the error correction rules are automatically extracted from a training corpus. Experimental results show that the proposed model improves 5.27% on the precision and 5.60% on the recall based on Eojoel unit for the speech recognition result.

• Journal of Korea Multimedia Society
• /
• 제9권1호
• /
• pp.1-10
• /
• 2006

In document clustering methods, the k-means algorithm and the Hierarchical Alglomerative Clustering(HAC) are often used. The k-means algorithm has the advantage of a processing time and HAC has also the advantage of a precision of classification. But both methods have mutual drawbacks, a slow processing time and a low quality of classification for the k-means algorithm and the HAC, respectively. Also both methods have the serious problem which is to compute a document similarity whenever new document is inserted into a cluster. A main property of web resource is to accumulate an information by adding new documents frequently. Therefore, we propose a new method of transitive closure tree based on the HAC method which can improve a processing time for a document clustering, and also propose a superior incremental clustering method for an insertion of a new document and a deletion of a document contained in a cluster. The proposed method is compared with those existing algorithms on the basis of a pre챠sion, a recall, a F-Measure, and a processing time and we present the experimental results.

• Journal of Korea Multimedia Society
• /
• 제16권4호
• /
• pp.502-509
• /
• 2013

In Information age, The academic liberal art computer education course set up goals to promote computer literacy and develop the ability to cope with changes in information society and improve productivity and national competitiveness. In this paper, we analyze on discovering of decisive variable and satisfaction index to have a influence on computer education on university students. As a preprocessing course, the proposed method selects optimum variable using correlation based feature selection(CFS) of machine learning tool based on Java and we calculate weighted value for each variable and then, we generate the optimal variable using weighted value based on fuzzy decision making method. we proposed Fuzzy decision making method in analysis of the academic liberal art computer education satisfaction index data and checked the accuracy of the satisfaction evaluation by using recall and precision.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• 제15권4호
• /
• pp.51-59
• /
• 2016

Measurements of the automobile manufacturers are available anywhere and anytime, directly based on the criterion of failure is measured. The maintenance of high-precision production activities is direct evidence of the fact that competitive manufacturing activities are very important in determining the success of companies to recall defective starting from raw material costs. The current manufacturing sites produce calipers and clearance gauge the degree of tool only specific. Therefore, judging the quality, including the number of errors, requires a lot of attention to the dimension failures in day-to-day measurements and measurement tasks and duties repeated in difficult situations. In this paper, we aim to develop a vehicle manufacturing plant site using each of the manufacturing processes while operating a measurement tool. We display it using the Image Processing PC-based S/W with all those visual facts by management and recorded as image information a more accurate and current situation to obtain information and share visual measurements. We carry out research on the design and development vision inspection algorithm applied for pattern-recognition techniques that can help manufacturing site quality control.