• ETRI Journal
• /
• 제42권2호
• /
• pp.230-238
• /
• 2020

Human-object interaction (HOI) detection is a popular computer vision task that detects interactions between humans and objects. This task can be useful in many applications that require a deeper understanding of semantic scenes. Current HOI detection networks typically consist of a feature extractor followed by detection layers comprising small filters (eg, 1 × 1 or 3 × 3). Although small filters can capture local spatial features with a few parameters, they fail to capture larger context information relevant for recognizing interactions between humans and distant objects owing to their small receptive regions. Hence, we herein propose a three-stream HOI detection network that employs a context convolution module (CCM) in each stream branch. The CCM can capture larger contexts from input feature maps by adopting combinations of large separable convolution layers and residual-based convolution layers without increasing the number of parameters by using fewer large separable filters. We evaluate our HOI detection method using two benchmark datasets, V-COCO and HICO-DET, and demonstrate its state-of-the-art performance.

• ETRI Journal
• /
• 제30권3호
• /
• pp.486-488
• /
• 2008

In this paper, a scheme to increase the throughput of RFID systems is presented, which considers the capture effect in the context of framed ALOHA protocol. Under the capture model in which the probability of one tag is identified successfully depending on the number of tags involved in the collision, two probabilistic methods for estimating the unknown number of tags are proposed. The first method is the maximum likelihood estimation method, and the second method is an approximate algorithm for reducing the computational time. The optimal frame size condition to maximize the system throughput by considering the capture effect is also presented.

• 대한임베디드공학회논문지
• /
• 제9권1호
• /
• pp.1-9
• /
• 2014

As Android platform and its applications are widely used, it is necessary to assure the reliability and safety of Android applications. The capture and replay technique is widely used for reproducing the crashes in Java applications. In this paper, we adapt the context-based capture and replay technique to Android applications by considering Android activity structure and its GUI characteristics. Our capture and replay technique uses probe code insertion based on AspectJ for minimizing the original code change and for enhancing the efficiency and reusability of the probe codes. For proving the efficiency of our technique, we provide a case study and execution overhead comparisons with related works, in which our approach has merits in GUI oriented Android applications with relatively low execution overhead.

• 한국정보전자통신기술학회논문지
• /
• 제15권4호
• /
• pp.249-256
• /
• 2022

드론 사진 측량의 경우 통상 상용 전문 SW인 Metashape, Pix4D Mapper, ContextCapture 및 간이 SW인 Global Mapper GIS의 "픽셀 to 포인트 도구" 모듈 등을 널리 사용하고 있다. 각 SW마다 고유의 항공 삼각측량법 해석에 대한 로직을 보유하고 있지만, 사용자가 SW를 선택하기 위해서는 지형공간정보의 좌표 값에 대한 비교 분석이 필요하다. 이를 위하여 드론 사진 측량을 위한 항공사진을 촬영하고, VRS-GPS 측량을 통하여 GCP 기준점 측량을 하여, 취득된 기초 데이터를 각 SW를 이용하여 데이터 처리를 한 후 정사 이미지과 DSM을 구축하고, GCP 기준점 측량 성과와 각 SW에서 취득된 정사 이미지 상의 GCP 대공표지의 중심점의 좌표(X,Y)및 DSM에 의한 GCP점의 높이 값(EL)을 비교했다. "공공측량 작업규정"에 따르면 각 SW의 결과치는 모두 오차범위 이내에 포함되어 어느 SW를 사용하더라도 규정에는 문제가 없는 것으로 판명되었다.

• 한국멀티미디어학회논문지
• /
• 제10권6호
• /
• pp.786-798
• /
• 2007

As we go about our daily lives, people often collect surrounding information and adapt to the situation. Computer development trends show that one wants computers to work like human beings, i.e. computers can sense its context and adapt corresponding to context changes. To implement this expectation, a context aware service layer is needed. In this layer, sensors capture its environment and send this information to the service center. Considering received information as its context, the service center seeks the suitable operation according to the context. Tills paper presents a context aware service which is applied in controlling air-conditioner. The air-conditioner includes sensors which are installed at some special positions in a room. Each of these sensors gathers comfort-influenced information like temperature, humidity and sends them to air-conditioner. The air-conditioner adapts its operation to the environment according to the sensed information. To control the air-conditioner effectively, we use a genetic algorithm which is suitable in adaptation issues. The simulation shows that the room condition can be maintained at a comfortable level by using context-aware services in the operation of the air-conditioning system.

• International journal of advanced smart convergence
• /
• 제13권3호
• /
• pp.176-182
• /
• 2024

Looking at the recent game market, classic games released in the past are being re-released with high-quality visuals, and users are generally satisfied. It can be said that the realization of realistic digital actors, which was not possible in the past, is now becoming a reality. Epic Games launched the MetaHuman Creator website in September 2021, allowing anyone to easily create realistic human characters. Since then, the number of animations created using MetaHumans has been increasing. As the characters become more realistic, the movement and expression animations expected by the audience must also be convincingly realized. Until recently, traditional methods were the primary approach for producing realistic character animations. For facial animation, Epic Games introduced an improved method on the Live Link app in 2023, which provides the highest quality among mobile-based techniques. In this context, this paper compares the results of animation produced using both keyframe facial capture and mobile-based capture. After creating an emotional expression animation with four sentences, the results were compared using Unreal Engine. While the facial capture method is more natural and easier to use, the precise and exaggerated expressions possible with the keyframe method cannot be overlooked, suggesting that a hybrid approach using both methods will likely continue for the foreseeable future.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권5호
• /
• pp.1396-1412
• /
• 2023

Conversation modeling is an important and challenging task in the field of natural language processing because it is a key component promoting the development of automated humanmachine conversation. Most recent research concerning conversation modeling focuses only on the current utterance (considered as the current question) to generate a response, and thus fails to capture the conversation's logic from its beginning. Some studies concatenate the current question with previous conversation sentences and use it as input for response generation. Another approach is to use an encoder to store all previous utterances. Each time a new question is encountered, the encoder is updated and used to generate the response. Our approach in this paper differs from previous studies in that we explicitly separate the encoding of the question from the encoding of its context. This results in different encoding models for the question and the context, capturing the specificity of each. In this way, we have access to the entire context when generating the response. To this end, we propose a deep neural network-based model, called the Context Model, to encode previous utterances' information and combine it with the current question. This approach satisfies the need for context information while keeping the different roles of the current question and its context separate while generating a response. We investigate two approaches for representing the context: Long short-term memory and Convolutional neural network. Experiments show that our Context Model outperforms a baseline model on both ConvAI2 Dataset and a collected dataset of conversational English.

• 한국통신학회논문지
• /
• 제17권2호
• /
• pp.99-113
• /
• 1992

다중접속 통신 방식으로 시분할 방식(TDMA)및 주파수분할 방식(FDMA)등이 보편적으로 사용되어 왔다. 그러나 초소형 위성 지구국(VSAT)을 사용한 위성통신, LAN을 사용한 데이터 통신등에 알로하(ALOHA)프로토콜 및 캐리어 감지 다중접속(CSMA)프로토콜등이 새로운 방식으로 등장하게 되었다.이들 새로운 방식들은 신호의 충돌현상(Collision)으로 시스템의 효율(Throughput)을 저하시키기는 하나 버스티 트래픽(Bursty Traffic)인 패킷 통신망에 적합한 통신 접속 기법임이 입증되어 왔다.본 연구는 무선 패킷 통신 시스템에서 페이딩(Fading)현상에 기인한 신호점유 효과 (Capture Effect)에 대한 통계적 고찰을 랜덤 다중접속 통신 시스템의 관점에서 조사하였다. 신호 점유현상은 계통의 처리량을 증가시키고 통신 지연을 줄일 수 있는 효과를 가지고 있다. 일반적인 페이딩 모델에 대하여 정확한 신호 점유확율을 도출하였고 다양한 페이딩 파라미터의 변화에 따른 수치 결과를 제시하였다. 또한 최소자승핏팅(Least-squares Fitting)방법을 사용하여 신호 점유 확률을 자수함수 형태로 근사화함으로써 알로하 통신 채널의 가용 처리 한계를 도출할 수 있었다.

• 대한공간정보학회지
• /
• 제5권2호
• /
• pp.55-63
• /
• 1997

문화재에 대한 정밀 수치자료의 획득과 공간분석 체계의 구축은 대상물의 보존 차원에서 매우 중요한 의미를 지니고 있다. 본 연구는 문화재즐 보다 효율적으로 관리하기 위한 기하학적 수치자료의 획득과 이의 정량적 공간분석에 관한 것이다. 공주 무령왕릉 군에 대한 동일좌표계 상에서의 3차원 정밀 실측과 영상해석을 실시하여 수치자료의 기반을 구축하고 다양한 기하학적 분석은 물론 3차원 모니터링을 수행하므로서 보존대책의 기초자료를 제공할 수 있었다.

• 한국경영정보학회:학술대회논문집
• /
• 한국경영정보학회 2007년도 추계학술대회
• /
• pp.273-277
• /
• 2007

Ubiquitous computing environment accelerates the advent of the IT Ecosystem. In the coming generation, a massive number of ubiquitous devices and services are converged into an ultra-large-scale system. In this context, high degrees of complexity and organization change the paradigm of knowledge and its management levels. The objective of this paper is to explore the Knowledge Management Systems in view of demands in ultra-large-scale systems. We introduce motivation for Next-Generation (NG) usage and their upcoming requirements. The possible applications will be discussed, and summary of different techniques are conducted. The tools and techniques that allow KMS to operate as "Vital Success Enabler" to achieve organizational benefits will be examined. Potential future directions for research are highlighted: these include advances in knowledge capture, storage / retrieval and sharing techniques, in particular with the surrounding role of information technology.