A5100393337- Advanced Data Storage Technologies
- Caching and Content Delivery
- Cloud Computing and Resource Management
- Parallel Computing and Optimization Techniques
- Advanced Vision and Imaging
- Distributed and Parallel Computing Systems
- Advanced Image and Video Retrieval Techniques
- Photonic and Optical Devices
- Green IT and Sustainability
- Video Surveillance and Tracking Methods
- Energy Efficient Wireless Sensor Networks
- Robotics and Sensor-Based Localization
- Educational Systems and Policies
- Face and Expression Recognition
- Optical measurement and interference techniques
- Advanced Fiber Optic Sensors
- Image Processing Techniques and Applications
- Industrial Vision Systems and Defect Detection
- Image and Video Stabilization
- Sensor Technology and Measurement Systems
- QR Code Applications and Technologies
- Technology and Data Analysis
- Face recognition and analysis
- Mobile Ad Hoc Networks
- Education and Learning Interventions
Samsung (South Korea)
2024
Korea University
2002-2024
Ajou University
2019-2024
Korea Advanced Institute of Science and Technology
2003-2022
Lutron Electronics (United States)
2022
Ulsan National Institute of Science and Technology
2019-2020
Virginia Tech
2017-2019
Hankyong National University
2002-2017
Sungkyunkwan University
2015
Qualcomm (United States)
2014
Users may face challenges while designing graphical user interfaces, due to a lack of relevant experience and guidance. This paper aims investigate the issues users during design process, how resolve them. To this end, we conducted semi-structured interviews, based on which built GUI prototyping assistance tool called GUIComp. can be connected software as an extension, it provides real-time, multi-faceted feedback user's current design. Additionally, two studies, in asked participants create...
Annotating tens or hundreds of tiny objects in a given image is laborious yet crucial for multitude Computer Vision tasks. Such imagery typically contains from various categories, the multi-class interactive annotation setting detection task has thus far been unex-plored. To address these needs, we propose novel method multiple instances classes, based on few point-based user in-puts. Our approach, C3Det, relates full context with annotator inputs local and global manner via late-fusion...
This paper proposes an object oriented face detection method using range and color information. Objects are segmented from the background stereo disparity histogram that represents information of objects. A matching pixel count (MPC) measure is introduced to enhance accuracy, removes effect unexpected noise in boundary region. For high-performance implementation MPC histogram, redundancy operations inherent area-based search operation removed. To detect facial regions among objects, a...
Key-value SSDs (KVSSDs) represent a major shift in the storage stack design, with numerous potential benefits. Despite this, their lack of native features critical to operation real world scenarios hinders adoption, and these benefits go unrealized. Moreover, simply adapting existing key-value stores run on KVSSDs proves underwhelming, as operate at lower raw device performance when compared modern block SSDs. This paper introduces Dotori. Dotori is KVSSD based store that provides much...
Existing conversational approaches for Internet of Things (IoT) service mashup do not support modification because the usability challenge, although it is common users to modify mashups in IoT environments. To through interfaces a usable manner, we propose agent (CoMMA). Users can using CoMMA natural language conversations. has two-step interaction, an implicature-based localization step, and step with disambiguation strategy. The allows easily search by vocalizing their expressions...
A novel method is proposed to improve the voice recognition performance by suppressing acoustic interferences that add nonlinear distortion a target recording signal when received device. The expected provide best in smart TV environments, where remote control collects command speech internal microphone and performs automatic recognition, secondary equipped set provides reference for background noise source. Due transmission channel, original interference corrupted nonlinearly, conventional...
There has been a constant demand for memory in modern mobile systems to provide users with better experience. Swapping is one of the cost-effective software solutions extra usable by reclaiming inactive pages and improving utilization. However, swapping not actively adopted since it incurs significant amount I/O, which fact impairs system performance as well user In this paper, we propose novel scheme properly harness systems. We identify that vast I/O comes from conflict traditional...
The conventional reflection-type extrinsic Fabry-Perot interferometric (EFPI) optical fiber sensor has good sensitivity and resolution compared with other types of sensors. However, they have the disadvantage that distinction strain direction EFPI is difficult because measurement method by only fringe counting. This paper presents newly developed transmission-type (TEFPI) sensor, which been improved additional functions, whose measuring system differs from sensors using a single-mode (SMF)...
The notion of object-based storage devices (OSDs) has been proposed to overcome the limitations traditional block-level interface which hinders development intelligent devices. main idea OSD is virtualize physical into a pool objects and offload burden space management device. We explore possibility adopting this for solid state drives (SSDs). SSDs (OSSDs) allow more efficient underlying flash storage, by utilizing object-aware data placement, hot/cold separation, QoS support prioritized...
As the mobile computing environment evolves, users demand high-quality apps and better user experience. Consequently, memory in devices has soared. Device manufacturers have fulfilled by equipping with more RAM. However, such a hardware approach is only temporary solution does not scale well resource-constrained environment. Meanwhile, systems adopt new app life cycle reclamation scheme tailored for cycle. When leaves an app, terminated but cached as long there enough free memory. If gets...
During the past decade, density of NAND flash memory has been increased in many folds. The increase driven by storing multiple bits a cell and scaling down fabrication process. Such advance manufacturing technology, however, significantly impaired reliability so that it becomes one major concerns use memory. Moreover, as writes data unit page, trend page size worsens amplifying small update to full programming. In this paper, we propose new programming method improve endurance cycle,...
Abstract Recently, embedded systems have become popular because of the rising demand for portable, low-power devices. A common task these devices is object tracking, which an essential part various applications. Until now, tracking in video sequences remains a challenging problem visual properties objects and their surrounding environments. Among approaches, particle filter has been proven effective dealing with difficulties tracking. In this research, we develop based method using color...
Since the adoption of hardware-accelerated features (e.g., hardware codec) improves performance and quality mobile devices, it revives need for contiguous memory allocation. However, physical in systems is highly fragmented due to frequent spawn exit processes lack proactive anti-fragmentation scheme. As a result, allocation large I/O buffers suffer from memory, thereby incurring high CPU usage power consumption. This paper presents approach that groups pages with same lifetime, stores them...
During the past decade, density of NAND flash memory has been increased in many folds. The increase driven by storing multiple bits a cell and scaling down fabrication process. Such advance manufacturing technology, however, significantly impaired reliability so that it becomes one major concerns use memory. Moreover, as writes data unit page, trend page size worsens amplifying small update to full programming. In this paper, we propose new programming method improve endurance cycle,...
GDDR memory has consistently maintained a leading position in delivering high performance: necessary for applications such as artificial intelligence, deep learning, and data centers. However, achieving an elevated I/O bandwidth, beyond the latest 27Gb/s GDDR6 [1], presents significant challenges ensuring sufficient link budget. One solution is using multi-level pulse amplitude modulation (PAM) signaling. In GDDR7 specification, single-ended PAM3 signaling used to achieve higher rates it can...
In this paper, we study the bandwidth efficiency (throughput) and energy of relay networks considering both physical layer medium access control (MAC) layer. Due to wireless signal power attenuation with transmission distance, using a for packet transmissions can lead more energy-efficient networking at expense requiring multihop transmissions. To understand potential benefits relay, energy-throughput tradeoff needs be analyzed. decentralized network, not only but also MAC should considered....
Abstract Graphene has been shown to be a prospective platform for active photonic devices with exotic optical properties. Employed in these devices, the graphene photodoping mechanism would allow remote spatiotemporal doping control by means of illumination, not restricted physical gate electrodes. This paper reports on efficient graphene‐CH 3 NH PbI perovskite heterostructure recently spotlighted photodetector applications. To maximize photoresponse, is optimized systematically introducing...
We introduce SOLAR 10.7B, a large language model (LLM) with 10.7 billion parameters, demonstrating superior performance in various natural processing (NLP) tasks. Inspired by recent efforts to efficiently up-scale LLMs, we present method for scaling LLMs called depth up-scaling (DUS), which encompasses depthwise and continued pretraining. In contrast other LLM methods that use mixture-of-experts, DUS does not require complex changes train inference efficiently. show experimentally is simple...
Due to the large measurement range, distributed optical fiber sensors are very effective in monitoring of structures. Brillouin sensors, which measure strain by using frequency shift, show some error for non-uniformly because their spatial resolution characteristics. They require compensation measured values and positions end-parts range not over all range. In this paper, we present characteristics method A deflection a beam subjected bending load is also presented. This verified three-point...