• ETRI Journal
• /
• v.46 no.2
• /
• pp.277-289
• /
• 2024

In this paper, we describe a neural network-based application that recommends multiple items using dialog context input and simultaneously outputs a response sentence. Further, we describe a multi-item recommendation by specifying it as a set of clothing recommendations. For this, a multimodal fusion approach that can process both cloth-related text and images is required. We also examine achieving the requirements of downstream models using a pretrained language model. Moreover, we propose a gate-based multimodal fusion and multiprompt learning based on a pretrained language model. Specifically, we propose an automatic evaluation technique to solve the one-to-many mapping problem of multi-item recommendations. A fashion-domain multimodal dataset based on Koreans is constructed and tested. Various experimental environment settings are verified using an automatic evaluation method. The results show that our proposed method can be used to obtain confidence scores for multi-item recommendation results, which is different from traditional accuracy evaluation.

• Bulletin of the Korean Mathematical Society
• /
• v.60 no.3
• /
• pp.785-827
• /
• 2023

In this paper, under some new appropriate conditions imposed on the parameter and mappings involved in the resolvent operator associated with a P-accretive mapping, its Lipschitz continuity is proved and an estimate of its Lipschitz constant is computed. This paper is also concerned with the construction of a new iterative algorithm using the resolvent operator technique and Nadler's technique for solving a new system of generalized multi-valued resolvent equations in a Banach space setting. The convergence analysis of the sequences generated by our proposed iterative algorithm under some appropriate conditions is studied. The final section deals with the investigation and analysis of the notion of H(·, ·)-co-accretive mapping which has been recently introduced and studied in the literature. We verify that under the conditions considered in the literature, every H(·, ·)-co-accretive mapping is actually P-accretive and is not a new one. In the meanwhile, some important comments on H(·, ·)-co-accretive mappings and the results related to them appeared in the literature are pointed out.

• Journal of Digital Convergence
• /
• v.10 no.8
• /
• pp.239-247
• /
• 2012

This paper is related to an Object VR-based virtual textile wearing system carrying out textile texture mapping based on viewpoint vector estimation and intensity difference map. The proposed system is characterized as capable of virtually wearing a new textile pattern selected by the user to the clothing shape section segmented from multi-view 2D images of clothes model for Object VR(Object Virtual Reality), and three-dimensionally viewing its virtual wearing appearance at multi-view points of the object. Regardless of color or intensity of model clothes, the proposed system is possible to virtually change the textile pattern with holding the properties of the selected clothing shape section, and also to quickly and easily simulate, compare, and select multiple textile pattern combinations for individual styles or entire outfits. The proposed system can provide higher practicality and easy-to-use interface, as it makes real-time processing possible in various digital environment, and creates comparatively natural and realistic virtual wearing styles, and also makes semi-automatic processing possible to reduce the manual works.

• Journal of Convergence for Information Technology
• /
• v.11 no.2
• /
• pp.25-30
• /
• 2021

In this paper, we studied the coding technique for flicker mitigation and BER performance improvement in visible light communication system. In order to increase the transmission speed of visible light communication, a multi-transmission multi-LED transmission system using a plurality of LEDs is being actively studied. However, when data is transmitted through N LEDs in a multi-LED visible light communication system using N LEDs, there is a continuous zero section in which 0 is transmitted simultaneously according to the data sequence, and since the transmission section of 1 is different, flickering Or, a phenomenon in which the dimming level changes occurs. The visible light communication system is a communication system that simultaneously performs communication and lighting functions. Therefore, transmission efficiency of communication and brightness of lighting must be considered at the same time. To solve this problem, we proposed a flicker reduction mapping that can alleviate flicker and dimming level problems, improve error detection and BER performance through coding mapping of each LED data sequence.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2006.06a
• /
• pp.68-71
• /
• 2006

Traditional section-based method develops blank along section planes and find trimming line by generating loop of end points. This method suffers from inaccurate results for regions with out-of-section motion. In this study, new fast method to find feasible trimming line is proposed. One step FEM is used to analyze the flanging and incremental development method is proposed to handle bad-shaped mesh and undercut part. Also in order to remedy mesh distortion during development, energy minimization technique is utilized. The proposed method is verified by shrink/stretch flange forming and successfully applied to the complex industrial applications such as door outer flanging process.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.3
• /
• pp.551-562
• /
• 2024

This study relates to a method and device for extracting and tracking moving objects. In particular, objects are extracted using different images between adjacent images, and the location information of the extracted object is continuously transmitted to provide accurate location information of at least one moving object. It relates to a method and device for extracting and tracking moving objects based on tracking moving objects. People tracking, which started as an expression of the interaction between people and computers, is used in many application fields such as robot learning, object counting, and surveillance systems. In particular, in the field of security systems, cameras are used to recognize and track people to automatically detect illegal activities. The importance of developing a surveillance system, that can detect, is increasing day by day.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.5 no.1
• /
• pp.7-17
• /
• 2010

ITS(Intelligent Transport System) as things are used for Broadcast service using TDMB/TPEG/NAVI rather than personal seamless service. It is attaching weight to Traffic information gathering, Charging, Settlement service. This research is applied to improve DCCP(Datagram Congestion Control Protocol) which has function as protecting data and preserving message boundary. The improving method is like that we solve data trust in UDP because Connection and Transmission overhead in UDP is less than in TCP. We fix the data loss which is generated from unordered delivery section of IP base wireless service by using DCCP protocol. We guarantee of connection with OBE(On-Board Equipment) and reliance about transmission of data by complement to mapping table and multi-hoping. Finally, We evaluate the performance about transmission of IP based data. We constructed a test-bed near research center for this test.

• Proceedings of the KOSOMBE Conference
• /
• v.1996 no.11
• /
• pp.142-146
• /
• 1996

In this paper, we present the main research results on the 2 Tesla class - superconducting MRI magnet which we have developed. Multi section type superconducting MRI main coil and various superconducting shims were designed and fabricated for obtaining the high field homogeneity, which is requested in the MR imaging. After assembling the magnet with room temperature bore cryostat field homogenity has been measured and analyzed by NMR field mapping system. According to this, field homogeneity of 22 ppm / 30 cm dsv was confirmed.

• Investigative Magnetic Resonance Imaging
• /
• v.11 no.2
• /
• pp.95-102
• /
• 2007

Purpose: To acquire high-resolution spiral-scan images at higher magnetic field, high homogeneous magnetic field is needed. Field inhomogeneity mapping and in-vivo shimming are important for rapid imaging such as spiral-scan imaging. The rapid scanning sequences are very susceptible to inhomogeneity. In this paper, we proposed a higher-order shimming method to obtain homogeneous magnetic field. Materials and Methods: To reduce measurement time for field inhomogeneity mapping, simultaneous axial/ sagittal, and coronal acquisitions are done using multi-slice based Fast Spin echo sequence. Acquired field inhomogeneity map is analyzed using the spherical harmonic functions, and shim currents are obtained by the multiplication of the pseudo-inverse of the field pattern with the inhomogeneity map. Results: Since the field inhomogeneity is increasing in proportion to the magnetic field, higher order shimming to reduce the inhomogeneity becomes more important in high field imaging. The shimming technique in which axial, sagittal, and coronal section inhomogeneity maps are obtained in one scan is developed, and the shimming method based on the analysis of spherical harmonics of the imhomogenity map is applied. The proposed technique is applicable to a localized shimming as well. High resolution spiral-scan imaging was successfully obtained with the proposed higher order shimming. Conclusion: Proposed pulse sequence for rapid measurement of inhomogeneity map and higher order shimming based on the inhomogeneity map work very well at 3 Tesla MRI system. With the proposed higher order shimming and localized higher order shimming techniques, high resolution spiral-scan images are successfully obtained at 3 T MRI system.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.5
• /
• pp.186-201
• /
• 2021

As autonomous driving technology advances, the accuracy of the vehicle position is important for recognizing the environments around driving. Map-matching localization techniques based on high definition (HD) maps have been studied to improve localization accuracy. Because conventional map-matching techniques estimate the vehicle position based on an HD map reference dataset representing the center of the lane, the estimated position does not reflect the deviation of the lateral distance within the lane. Therefore, this paper proposes a localization system based on the reference lateral position dataset extracted using image processing and HD maps. Image processing extracts the driving lane number using inverse perspective mapping, multi-lane detection, and yellow central lane detection. The lane departure method estimates the lateral distance within the lane. To collect the lateral position reference dataset, this approach involves two processes: (i) the link and lane node is extracted based on the lane number obtained from image processing and position from GNSS/INS, and (ii) the lateral position is matched with the extracted link and lane node. Finally, the vehicle position is estimated by matching the GNSS/INS local trajectory and the reference lateral position dataset. The performance of the proposed method was evaluated by experiments carried out on a highway environment. It was confirmed that the proposed method improves accuracy by about 1.0m compared to GNSS / INS, and improves accuracy by about 0.04m~0.21m (7~30%) for each section when compared with the existing lane-level map matching method.