• Investigative Magnetic Resonance Imaging
• /
• v.25 no.4
• /
• pp.252-265
• /
• 2021

Cardiac magnetic resonance imaging (MRI) serves as a clinical gold-standard non-invasive imaging technique for the assessment of global and regional cardiac function. Conventional cardiac MRI is limited by the long acquisition time, the need for ECG gating and/or long breathhold, and insufficient spatiotemporal resolution. Real-time cardiac cine MRI refers to high spatiotemporal cardiac imaging using data acquired continuously without synchronization or binning, and therefore of potential interest in overcoming the limitations of conventional cardiac MRI. Novel acquisition and reconstruction techniques must be employed to facilitate real-time cardiac MRI. The goal of this study is to discuss methods that have been developed for real-time cardiac MRI. In particular, we classified existing techniques into two categories based on the use of non-iterative and iterative reconstruction. In addition, we present several research trends in this direction, including deep learning-based image reconstruction and other advanced real-time cardiac MRI strategies that reconstruct images acquired from real-time free-breathing techniques.

• International Journal of Control, Automation, and Systems
• /
• v.5 no.6
• /
• pp.674-680
• /
• 2007

The probabilistic roadmap (PRM) method, which is a popular path planning scheme, for a manipulator, can find a collision-free path by connecting the start and goal poses through a roadmap constructed by drawing random nodes in the free configuration space. PRM exhibits robust performance for static environments, but its performance is poor for dynamic environments. On the other hand, reinforcement learning, a behavior-based control technique, can deal with uncertainties in the environment. The reinforcement learning agent can establish a policy that maximizes the sum of rewards by selecting the optimal actions in any state through iterative interactions with the environment. In this paper, we propose efficient real-time path planning by combining PRM and reinforcement learning to deal with uncertain dynamic environments and similar environments. A series of experiments demonstrate that the proposed hybrid path planner can generate a collision-free path even for dynamic environments in which objects block the pre-planned global path. It is also shown that the hybrid path planner can adapt to the similar, previously learned environments without significant additional learning.

• Journal of the Korean Society of Industry Convergence
• /
• v.27 no.1
• /
• pp.87-95
• /
• 2024

This study is on modulo scheduling algorithms for multicore processor in machine learning applications. Machine learning algorithms are designed to perform a large amount of operations such as vectors and matrices in order to quickly process large amounts of data stream. To support such large amounts of computations, processor architectures to support applications such as artificial intelligence, neural networks, and machine learning are designed in the form of parallel processing such as multicore. To effectively utilize these multi-core hardware resources, various compiler techniques are being used and studied. In this study, among these compiler techniques, we analyzed the modular scheduler, which is especially important in one core's computation pipeline. This paper looked at and compared the iterative modular scheduler and the swing modular scheduler, which are the most widely used and studied. As a result, both schedulers provided similar performance results, and when measuring register pressure as an indicator, it was confirmed that the swing modulo scheduler provided slightly better performance. In this study, a technique that divides recurrence edge is proposed to improve the minimum initiation interval of the modulo schedulers.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.169-175
• /
• 2021

Recently, an ultrasound hologram and its applications have gained attention in the ultrasound research field. However, the determination technique of transmit signal phases, which generate a hologram, has not been significantly advanced from the previous algorithms which are time-consuming iterative methods. Thus, we applied the deep learning technique, which has been previously adopted to generate an optical hologram, to generate an ultrasound hologram. We further examined the Deep learning-based Holographic Ultrasound Generation algorithm (Deep-HUG). We implement the U-Net-based algorithm and examine its generalizability by training on a dataset, which consists of randomly distributed disks, and testing on the alphabets (A-Z). Furthermore, we compare the Deep-HUG with the previous algorithm in terms of computation time, accuracy, and uniformity. It was found that the accuracy and uniformity of the Deep-HUG are somewhat lower than those of the previous algorithm whereas the computation time is 190 times faster than that of the previous algorithm, demonstrating that Deep-HUG has potential as a useful technique to rapidly generate an ultrasound hologram for various applications.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.2
• /
• pp.244-251
• /
• 2007

Intention reading technique is essential to provide personalized services toward more convenient and human-friendly services in complex ubiquitous environment such as a smart home. If a system has knowledge about an user's intention of his/her behavioral pattern, the system can provide mote qualified and satisfactory services automatically in advance to the user's explicit command. In this sense, learning capability is considered as a key function for the intention reading technique in view of knowledge discovery. In this paper, ore introduce a personalized media control method for a possible application iii a smart home. Note that data pattern such as human behavior contains lots of inconsistent data due to limitation of feature extraction and insufficiently available features, where separable data groups are intermingled with inseparable data groups. To deal with such a data pattern, we introduce an effective engineering approach with the combination of fuzzy logic and probabilistic reasoning. The proposed learning system, which is based on IFCS (Iterative Fuzzy Clustering with Supervision) algorithm, extract probabilistic fuzzy rules effectively from the given numerical training data pattern. Furthermore, an extended architectural design methodology of the learning system incorporating with the IFCS algorithm are introduced. Finally, experimental results of the media contents recommendation system are given to show the effectiveness of the proposed system.

• Nuclear Engineering and Technology
• /
• v.53 no.7
• /
• pp.2341-2347
• /
• 2021

Because single-photon emission computed tomography (SPECT) is one of the widely used nuclear medicine imaging systems, it is extremely important to acquire high-quality images for diagnosis. In this study, we designed a super-resolution (SR) technique using dense block-based deep convolutional neural network (CNN) and evaluated the algorithm on real SPECT phantom images. To acquire the phantom images, a real SPECT system using a^99mTc source and two physical phantoms was used. To confirm the image quality, the noise properties and visual quality metric evaluation parameters were calculated. The results demonstrate that our proposed method delivers a more valid SR improvement by using dense block-based deep CNNs as compared to conventional reconstruction techniques. In particular, when the proposed method was used, the quantitative performance was improved from 1.2 to 5.0 times compared to the result of using the conventional iterative reconstruction. Here, we confirmed the effects on the image quality of the resulting SR image, and our proposed technique was shown to be effective for nuclear medicine imaging.

• Journal of the Korean Society of Radiology
• /
• v.17 no.1
• /
• pp.37-46
• /
• 2023

This study aims to evaluate the usefulness of Deep Learning Image Reconstruction (TrueFidelity, TF), the image quality of existing Filtered Back Projection (FBP) and Adaptive Statistical Iterative Reconstruction-Veo (ASIR-V) were compared. Noise, CNR, and SSIM were measured by obtaining images with doses fixed at 17.29 mGy and altered to 10.37 mGy, 12.10 mGy, 13.83 mGy, and 15.56 mGy in reconstruction techniques of FBP, ASIR-V 50%, and TF-H. TF-H has superior image quality compared to FBP and ASIR-V when the reconstruction technique change is given at 17.29 mGy. When dose changes were made, Noise, CNR, and SSIM were significantly different when comparing 10.37 mGy TF-H and FBP (p<0.05), and no significant difference when comparing 10.37 mGy TF-H and ASIR-V 50% (p>0.05). TF-H has a dose-reduction effect of 30%, as the highest dose of 15.56 mGy ASIR-V has the same image quality as the lowest dose of 10.37 mGy TF-H. Thus, Deep Learning Reconstruction techniques (TF) were able to reduce dose compared to Iterative Reconstruction techniques (ASIR-V) and Filtered Back Projection (FBP). Therefore, it is considered to reduce the exposure dose of patients.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.9
• /
• pp.35-42
• /
• 1998

Finding a technique to achieve high machining precision and high productivity is an important issue for CNC machining. One of the solutions to meet better performance of machining is feedrate control. In this paper we present an adaptive feedrate neuro-control method for high precision and high speed machining. The adaptive neuro-control architecture consists of a neural network identifier(NNI) and an iterative learning control algorithm with inversion of the NNI. The NNI is an identifier for the nonlinear characteristics of feedrate and contour error, which is utilized in iterative learning for adaptive feedrate control with specified contour error tolerance. The proposed neuro-control method has been successfully evaluated for machining circular, corner and involute contours by computer simulations.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.10
• /
• pp.11-21
• /
• 2018

In this paper we propose a novel log template discovery algorithm which achieves high quality of discovered log templates through iterative log filtering technique. Log templates are the static string pattern of logs that are used to produce actual logs by inserting variable values during runtime. Identifying individual logs into their template category correctly enables us to conduct automated analysis using state-of-the-art machine learning techniques. Our technique looks at the group of logs column-wise and filters the logs that have the value of the highest proportion. We repeat this process per each column until we are left with highly homogeneous set of logs that most likely belong to the same log template category. Then, we determine which column is the static part and which is the variable part by vertically comparing all the logs in the group. This process repeats until we have discovered all the templates from given logs. Also, during this process we discover the custom patterns such as ID formats that are unique to the application. This information helps us quickly identify such strings in the logs as variable parts thereby further increasing the accuracy of the discovered log templates. Existing solutions suffer from log templates being too general or too specific because of the inability to detect custom patterns. Through extensive evaluations we have learned that our proposed method achieves 2 to 20 times better accuracy.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.6
• /
• pp.1867-1873
• /
• 2000

Function approximation from a set of input-output pairs has numerous applications in scientific and engineering areas. Support vector machine (SVM) is a new and very promising classification, regression and function approximation technique developed by Vapnik and his group at AT&TG Bell Laboratories. However, it has failed to establish itself as common machine learning tool. This is partly due to the fact that this is not easy to implement, and its standard implementation requires the use of optimization package for quadratic programming (QP). In this appear we present simple iterative Kernel Adatron (KA) algorithm for function approximation and compare it with standard SVM algorithm using QP.