• Tunnel and Underground Space
• /
• v.33 no.4
• /
• pp.281-298
• /
• 2023

Continuous research efforts are being devoted to unmanned mobile platforms for lunar exploration. There is an ongoing demand for real-time information processing to accurately determine the positioning and mapping of areas of interest on the lunar surface. To apply deep learning processing and analysis techniques to practical rovers, research on software integration and optimization is imperative. In this study, a foundational investigation has been conducted on real-time analysis of virtual lunar base construction site images, aimed at automatically quantifying spatial information of key objects. This study involved transitioning from an existing region-based object recognition algorithm to a boundary box-based algorithm, thus enhancing object recognition accuracy and inference speed. To facilitate extensive data-based object matching training, the Batch Hard Triplet Mining technique was introduced, and research was conducted to optimize both training and inference processes. Furthermore, an improved software system for object recognition and identical object matching was integrated, accompanied by the development of visualization software for the automatic matching of identical objects within input images. Leveraging satellite simulative captured video data for training objects and moving object-captured video data for inference, training and inference for identical object matching were successfully executed. The outcomes of this research suggest the feasibility of implementing 3D spatial information based on continuous-capture video data of mobile platforms and utilizing it for positioning objects within regions of interest. As a result, these findings are expected to contribute to the integration of an automated on-site system for video-based construction monitoring and control of significant target objects within future lunar base construction sites.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.3
• /
• pp.10-17
• /
• 2012

In general, image-based 3D scenes can now be found in many popular vision systems, computer games and virtual reality tours. In this paper, we propose a method for creating 3D virtual scenes based on 2D image that is completely automatic and requires only a single scene as input data. The proposed method is similar to the creation of a pop-up illustration in a children's book. In particular, to estimate geometric structure information for 3D scene from a single outdoor image, we apply the tensor voting to an image segmentation. The tensor voting is used based on the fact that homogeneous region in an image is usually close together on a smooth region and therefore the tokens corresponding to centers of these regions have high saliency values. And then, our algorithm labels regions of the input image into coarse categories: "ground", "sky", and "vertical". These labels are then used to "cut and fold" the image into a pop-up model using a set of simple assumptions. The experimental results show that our method successfully segments coarse regions in many complex natural scene images and can create a 3D pop-up model to infer the structure information based on the segmented region information.

• Journal of KIISE
• /
• v.43 no.3
• /
• pp.370-379
• /
• 2016

Recently personal media were produced in a variety of ways as a lot of smart devices have been spread and services using these data have been desired. Therefore, research has been actively conducted for the media analysis and recognition technology and we can recognize the meaningful object from the media. The system using the media ontology has the disadvantage that can't classify the media appearing in the video because of the use of a video title, tags, and script information. In this paper, we propose a system to automatically classify video using the objects shown in the media data. To do this, we use a description logic-based reasoning and a rule-based inference for event processing which may vary in order. Description logic-based reasoning system proposed in this paper represents the relation of the objects in the media as activity ontology. We describe how to another rule-based reasoning system defines an event according to the order of the inference activity and order based reasoning system automatically classify the appropriate event to the category. To evaluate the efficiency of the proposed approach, we conducted an experiment using the media data classified as a valid category by the analysis of the Youtube video.

• Journal of KIISE:Software and Applications
• /
• v.34 no.12
• /
• pp.1045-1055
• /
• 2007

Latent topic models are statistical models which automatically captures salient patterns or correlation among features underlying a data collection in a probabilistic way. They are gaining an increased popularity as an effective tool in the application of automatic semantic feature extraction from text corpus, multimedia data analysis including image data, and bioinformatics. Among the important issues for the effectiveness in the application of latent topic models to the massive data set is the efficient learning of the model. The paper proposes an accelerated learning technique for PLSA model, one of the popular latent topic models, by an incremental EM algorithm instead of conventional EM algorithm. The incremental EM algorithm can be characterized by the employment of a series of partial E-steps that are performed on the corresponding subsets of the entire data collection, unlike in the conventional EM algorithm where one batch E-step is done for the whole data set. By the replacement of a single batch E-M step with a series of partial E-steps and M-steps, the inference result for the previous data subset can be directly reflected to the next inference process, which can enhance the learning speed for the entire data set. The algorithm is advantageous also in that it is guaranteed to converge to a local maximum solution and can be easily implemented just with slight modification of the existing algorithm based on the conventional EM. We present the basic application of the incremental EM algorithm to the learning of PLSA and empirically evaluate the acceleration performance with several possible data partitioning methods for the practical application. The experimental results on a real-world news data set show that the proposed approach can accomplish a meaningful enhancement of the convergence rate in the learning of latent topic model. Additionally, we present an interesting result which supports a possible synergistic effect of the combination of incremental EM algorithm with parallel computing.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.2
• /
• pp.155-160
• /
• 2014

There are many image processing based algorithms and applications that fail when illumination change occurs. Therefore, the illumination change has to be detected then the illumination change occurred images need to be enhanced in order to keep the appropriate algorithm processing in a reality. In this paper, a new method for detecting illumination changes efficiently in a real time by using local region information and fuzzy logic is introduced. The effective way for detecting illumination changes in lighting area and the edge of the area was selected to analyze the mean and variance of the histogram of each area and to reflect the changing trends on previous frame's mean and variance for each area of the histogram. The ways are used as an input. The changes of mean and variance make different patterns w hen illumination change occurs. Fuzzy rules were defined based on the patterns of the input for detecting illumination changes. Proposed method was tested with different dataset through the evaluation metrics; in particular, the specificity, recall and precision showed high rates. An automatic parameter selection method was proposed for contrast limited adaptive histogram equalization method by using entropy of image through adaptive neural fuzzy inference system. The results showed that the contrast of images could be enhanced. The proposed algorithm is robust to detect global illumination change, and it is also computationally efficient in real applications.

• Journal of Biomedical Engineering Research
• /
• v.19 no.5
• /
• pp.477-486
• /
• 1998

In this paper, we implemented the urine analysis system capable of measuring a qualitative and semi-quantitative and assay using strip. The analysis algorithm of urine analysis was adopted a fuzzy logic-based classifiers that was robust to external error factors such as temperature and electric power noises. The spectroscopic properties of 9 pads In a strip were studied to developing the urine analysis system was designed for robustnesss and stability. The urine analysis system was consisted of hardware and software. The hardware of the urine analysis system was based on one-chip microprocessor, and Its peripherals which composed of optic modulo, tray control, preamplifier, communication with PC, thermal printer and operating status indicator. The software of the urine analysis system was composed of system program and classification program. The system program did duty fort system control, data acquisition and data analysis. The classification program was composed of fuzzy inference engine and membership function generator. The membership function generator made triangular membership functions by statical method for quality control. Resulted data was transferred through serial cable to PC. The transferred data was arranged and saved be data acquisition program coded by C+ + language. The precision of urine analysis system and the stability of fuzzy classifier were evaluated by testing the standard urine samples. Experimental results showed a good stability states and a exact classification.

• Journal of Information Technology Application
• /
• v.2 no.1
• /
• pp.1-24
• /
• 2000

Project management in the Construction field inherently has more uncertainty and more risks relative to ones from other area. This is the very reason for why project management is recognized as the important task to construction companies. For getting better performance in the project management, we need a system that keeps the consistencies in a automatic or semi-automatic manner through the project management stages like as project definition stage, project planning stage, project design and implementation stage. But since the early stages such as definition and planning stages has many unstructured features and also are dependent to unique expertise or experience of a specific company, we have difficulty providing systematic support for the task of these stages. This kind of problem becomes harder to solve especially in the plant construction domain that is our target domain. Therefore, in this paper, we propose and also implement a systematic approach to resolve the problem mentioned for the early project management stages in the plant construction domain. The results of our approach can be used not only for the purpose of the early project management stages but also can be used automatically as an input to commercial project management tools for the middle project management stages. Because of doing in this way, the construction project can be consistently managed from the definition to implementation stage in a seamless manner. For achieving this purpose, we adopt knowledge based inference, CBR, and neural network as major methodologies and we also applied our approach to two real world cases, power plant and drainage treatment plant cases from a leading construction company in Korea. Since these two application cases showed us very successful results, we can say our approach was validated successfully to the plant construction area. Finally, we believe our approach will contribute to many project management problems from more broader construction area.

• Journal of Broadcast Engineering
• /
• v.17 no.2
• /
• pp.270-283
• /
• 2012

With the advent of multi-channel TV, IPTV and smart TV services, excessive amounts of TV program contents become available at users' sides, which makes it very difficult for TV viewers to easily find and consume their preferred TV programs. Therefore, the service of automatic TV recommendation is an important issue for TV users for future intelligent TV services, which allows to improve access to their preferred TV contents. In this paper, we present a recommendation model based on statistical machine learning using a collaborative filtering concept by taking in account both public and personal preferences on TV program contents. For this, users' preference on TV programs is modeled as a latent topic variable using LDA (Latent Dirichlet Allocation) which is recently applied in various application domains. To apply LDA for TV recommendation appropriately, TV viewers's interested topics is regarded as latent topics in LDA, and asymmetric Dirichlet distribution is applied on the LDA which can reveal the diversity of the TV viewers' interests on topics based on the analysis of the real TV usage history data. The experimental results show that the proposed LDA based TV recommendation method yields average 66.5% with top 5 ranked TV programs in weekly recommendation, average 77.9% precision in bimonthly recommendation with top 5 ranked TV programs for the TV usage history data of similar taste user groups.

• Phonetics and Speech Sciences
• /
• v.15 no.3
• /
• pp.75-82
• /
• 2023

Automatic speech recognition (ASR) has been revolutionized with deep learning-based approaches, among which self-supervised learning methods have proven to be particularly effective. In this study, we aim to enhance the performance of OpenAI's Whisper model, a multilingual ASR system on the Korean language. Whisper was pretrained on a large corpus (around 680,000 hours) of web speech data and has demonstrated strong recognition performance for major languages. However, it faces challenges in recognizing languages such as Korean, which is not major language while training. We address this issue by fine-tuning the Whisper model with an additional dataset comprising about 1,000 hours of Korean speech. We also compare its performance against a Transformer model that was trained from scratch using the same dataset. Our results indicate that fine-tuning the Whisper model significantly improved its Korean speech recognition capabilities in terms of character error rate (CER). Specifically, the performance improved with increasing model size. However, the Whisper model's performance on English deteriorated post fine-tuning, emphasizing the need for further research to develop robust multilingual models. Our study demonstrates the potential of utilizing a fine-tuned Whisper model for Korean ASR applications. Future work will focus on multilingual recognition and optimization for real-time inference.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.7
• /
• pp.1319-1325
• /
• 2006

Urine analysis is basic test in clinical medicine using visual examination by expert nurse. Recently, this test is measured by automatic urine analysis system. But, this system has different results by each instrument. So, a new classification algorithm is required for accurate classify and urine color collection. In this paper, a intelligent color classifier of urine analysis system was designed using neural network algorithm. The input parameters are three stimulus(RGB) after preprocessing using normalization. The fuzzy inference and neural network ware constructed for classify class according to 9 urine test items and $3{\sim}7$ classes. The experiment material to be used a standard sample of medicine. The possibility to adapt classifier designed for urine analysis system was verified as classifying measured standard samples and observing classified result. Of many test items, experimental results showed a satisfactory agreement with test results of reference system.