• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.27-29
• /
• 2021

인공지능 기반 머신 비전 응용이 증가함에 따라 사람이 아닌 기계에서 소비되는 영상 정보를 전송하는 요구가 발생하고 있다. 일반적으로 영상 정보를 전송할 때는 전송 비용을 고려하여 정보를 압축하며 기존 영상 압축 방법은 사람의 시각 인지적 특성을 반영하여 설계되었다. 따라서 기존 영상 압축 방법은 기계에서 소비되는 영상 정보를 압축하는 방법으로 적절하지 않다고 판단하여 2019년 7월, 기계를 위한 영상 부호화 기술의 표준화가 시작되었다. 본 논문에서는 머신 비전 태스크 중, 객체 탐지를 수행하는 네트워크의 피처 맵을 압축하는 방법을 제안한다. 제안하는 방법은 피처 맵의 채널 간 중복성을 제거하기 위해 PCA 기반의 변환을 적용하여 피처 맵의 차원을 축소하며 특히 해상도 계층 구조를 갖는 네트워크의 피처 맵을 압축하기 위해 각 해상도 계층간 변환 기저를 예측하여 추가로 압축률을 높인다. 제안하는 방법을 적용하여 객체 탐지 결과의 큰 성능 하락 없이 약 92.3%에 데이터양 감소를 달성하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.140-142
• /
• 2022

최근 신경망 기반 기술들의 발달에 따라, 신경망 기술들은 충분히 높은 임무 수행 성능을 달성하고 있으며 사물인터넷, 스마트시티, 자율주행 등 다양한 환경을 고려한 응용 역시 활발히 연구되고 있다. 하지만 이러한 신경망의 임무 다양성과 복잡성은 더욱 많은 비디오 데이터가 요구되며 대역폭이 제한된 환경을 고려한 응용에서 이러한 비디오 데이터를 효과적으로 전송할 방법이 필요하다. 이에 따라 국제 표준화 단체인 MPEG 에서는 신경망 기계 소비에 적합한 비디오 부호화 표준 개발을 위해 Video Coding for Machines (VCM) 표준화를 진행하고 있다. 본 논문에서는 신경망의 특징 부호화 효율을 개선하기 위하여 VCM 을 위한 다중 스케일 특징 압축 방법을 제안한다. COCO2017 데이터셋의 검증 영상을 기반으로 제안방법을 평가한 결과, 압축된 특징의 크기는 원본 이미지의 0.03 배이며 6.8% 미만의 임무 정확도 손실을 보였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.4
• /
• pp.562-567
• /
• 2009

As a key mechanism of the human emotion interaction, Facial Expression is a powerful tools in HRI(Human Robot Interface) such as Human Computer Interface. By using a facial expression, we can bring out various reaction correspond to emotional state of user in HCI(Human Computer Interaction). Also it can infer that suitable services to supply user from service agents such as intelligent robot. In this article, We addresses the issue of expressive face modeling using an advanced active appearance model for facial emotion recognition. We consider the six universal emotional categories that are defined by Ekman. In human face, emotions are most widely represented with eyes and mouth expression. If we want to recognize the human's emotion from this facial image, we need to extract feature points such as Action Unit(AU) of Ekman. Active Appearance Model (AAM) is one of the commonly used methods for facial feature extraction and it can be applied to construct AU. Regarding the traditional AAM depends on the setting of the initial parameters of the model and this paper introduces a facial emotion recognizing method based on which is combined Advanced AAM with Bayesian Network. Firstly, we obtain the reconstructive parameters of the new gray-scale image by sample-based learning and use them to reconstruct the shape and texture of the new image and calculate the initial parameters of the AAM by the reconstructed facial model. Then reduce the distance error between the model and the target contour by adjusting the parameters of the model. Finally get the model which is matched with the facial feature outline after several iterations and use them to recognize the facial emotion by using Bayesian Network.

• Journal of the Korea Computer Graphics Society
• /
• v.22 no.3
• /
• pp.21-29
• /
• 2016

Cell segmentation is an important but time-consuming and laborious task in biological image analysis. An automated, robust, and fast method is required to overcome such burdensome processes. These needs are, however, challenging due to various cell shapes, intensity, and incomplete boundaries. A precise cell segmentation will allow to making a pathological diagnosis of tissue samples. A vast body of literature exists on cell segmentation in microscopy images [1]. The majority of existing work is based on input images and predefined feature models only - for example, using a deformable model to extract edge boundaries in the image. Only a handful of recent methods employ data-driven approaches, such as supervised learning. In this paper, we propose a novel data-driven cell segmentation algorithm for bright-field microscopy images. The proposed method minimizes an energy formula defined by two dictionaries - one is for input images and the other is for their manual segmentation results - and a common sparse code, which aims to find the pixel-level classification by deploying the learned dictionaries on new images. In contrast to deformable models, we do not need to know a prior knowledge of objects. We also employed convolutional sparse coding and Alternating Direction of Multiplier Method (ADMM) for fast dictionary learning and energy minimization. Unlike an existing method [1], our method trains both dictionaries concurrently, and is implemented using the GPU device for faster performance.

• Korean Journal of Microbiology
• /
• v.47 no.1
• /
• pp.14-21
• /
• 2011

Riboflavin synthase catalyzes the formation of one molecule of each riboflavin and 5-amino-6-ribitylamino-2,4-pyrimidinedione by the transfer of a 4-carbon moiety between two molecules of the substrates, 6,7-dimetyl-8-ribityllumazine. The most remarkable feature is the sequence similarity between the N-terminal half (1-97) and the C-terminal half domain (99-213). To investigate the structure and fluorescent characteristics of the N-terminal half of riboflavin synthase (N-RS) in Escherichia coli, more than 10 mutant genes coding for the mutated N-terminal domain of riboflavin synthase were generated by polymerase chain reaction. The genes coding for the proteins were inserted into pQE vector designed for easy purification of protein by 6X-His tagging system, expressed, and the proteins were purified. Almost all mutated N-terminal domain of riboflavin synthases bind to 6,7-dimethyl-8-ribityllumazine and riboflavin as fluorescent ligands. However, N-RS C47D and N-RS ET66,67DQ mutant proteins show colorless, indicating that fluorescent ligands were dissociated during purification. In addition, most mutated proteins show low fluorescent intensity comparing to N-RS wild type, whereas N-RS C48S posses stronger fluorescent intensity than that of wild type protein. Based on this result, N-RS C48S can be used as the tool for high throughput screening system for searching for the compound with inhibitory effect for the riboflavin synthase.

• Journal of Broadcast Engineering
• /
• v.22 no.1
• /
• pp.1-14
• /
• 2017

This paper proposes adaptive reference structure decision method to improve the performance of HEVC (High Efficiency Video Coding) encoder. When an event occurs in the input sequence, such as scene change, scene rotation, fade in/out, or light on/off, the proposed algorithm changes the reference structure to improve the inter prediction performance. The proposed algorithm divides GOP (Group Of Pictures) into two sub-groups based on the picture that has such event and decides the reference pictures in the divided sub-groups. Also, this paper proposes fast encoding method which changes the picture type of first encoded picture in the GOP that has such event to CRA (Clean Random Access). With the statistical feature that intra prediction is selected by high probability for the first encoded picture in the GOP carrying such event, the proposed fast encoding method does not operate inter prediction. The experimental result shows that the proposed adaptive reference structure decision method improves the BD-rate 0.3% and reduces encoding time 4.9% on average under the CTC (Common Test Condition) for standardization. In addition, the proposed reference structure decision method with the picture type change reduces the average encoding time 12.2% with 0.11% BD-rate loss.

• The Korean Nurse
• /
• v.37 no.2
• /
• pp.44-52
• /
• 1998

The purpose of this study is to inquire the people's views on nursing for nurses, correct the image of nurse and take it as basis to be applied on nursing education examining the image of nursing on Television drama playing important role of mass media. 22 nurses of the characters in drama is applied to the analysis object of this study by selecting 6 dramas of Television ones the nurse play on the prime time from June 1 to August 31 in 1997. Contents analysis method was used in Data Analysis, 4 items was used after Coders previously modify and compensate it based on research documents of 1m Milim(1996) 2 Coders made the Coding the article on each person by them seeing the recorded film making the Coding Paper each items is written by the character. The average of reliability degree was 90% which measured the reliability degree by the mathod of Holsti. The statisic method of frequency, percentage was used SPSS Program in data processing The results were as follows. 1. Relative importance of 86.2% nurses in drama was depicted as extra characters 2. The affair attitude of nurses shown on drama was revealed as mechanical(84.7%), passive(45.5%), dependent(54.4%) unkind(68.2%). 3. The activity of nurses was classified with professional! simple affair. The professional affairs such as I.V., Blood Pressure Check, Rounding, Nursing Recording, Patient Education, Assist of Operation, Assistant meal of Patient, etc is mainly depicted and the screen of simple affair such as Receiving telephone, Carrying Tray or Dragging, Stretcher Car, Dressing Car and or Wheel Chair than professional affair. 4. The appearance feature of nurses was shown on thin physique(68.2%), common stature(68.2), dirty costume(45.4%), common appearance(81.9%), unnoble action(63.6%). The image of nurses is illuminated as the exterial scene of technical affair such as assisting the doctors and affair focused on accident and educational activity of nureses or extended role is nor depicted on Television drama. Therefore, the people regard the nurse as sexual object with good appearance than professional worker working professional nursing We want the following, epigraph based on above conclusion. 1. The continuous research is required on the image of nurse shown on various mass media. 2. The later research is required on appliction strategy of mass media for advancing the image of nurse. 3. The research to strengthen the objectivity by comparing analyzed data on drama & analyzing it is required 4. Through the deep study, the standard to show a concrete and professional work of nurses to scenario writers of TV drama is suggested by the association. 5. The monitoring about the mass media must be activated, not by some nurses, on a national scale and much study on the basis of this is needed.

• The Journal of the Korea Contents Association
• /
• v.14 no.11
• /
• pp.362-369
• /
• 2014

This study aimed to propose an accurate diagnostic method for osteoporosis by realizing a computer-aided diagnosis system with the application of the statistical analysis of texture features using digital images of lateral lumbar spine of patients with osteoporosis and providing reliable supplementary diagnostic information by model experimental research for early diagnosis of diseases. For these purposes, digital images of lateral lumbar spine of normal individuals and patients with osteoporosis were used in the experiments, and the values of statistical texture features on the set ROI were expressed in six parameters. Among the texture feature values of the six parameters of osteoporosis, the highest and lowest recognition rates of 95 and 80% were shown in average gray level and uniformity, respectively. Moreover, all the six parameters showed recognition rates of over 80% for osteoporosis: 82.5% in average contrast, 90% in smoothness, 87.5% in skewness, and 87.5% in entropy. Therefore, if a program developing into a computer-aided diagnosis system for medical images is coded based on the results of this study, it is considered possible to be applied to preliminary diagnostic data for automatic detection of lesions and disease diagnosis using medical images, to provide information for definite diagnosis of diseases, to diagnose by limited device, and to be used to shorten the time to analyze medical images.

• The KIPS Transactions:PartD
• /
• v.19D no.3
• /
• pp.229-236
• /
• 2012

The demand for map applications in both Web and mobile environments has been rapidly increased with the population of Web and smart phone usage. Web-based map applications are mostly developed on such environments as ArcGIS and MapServer and mobile map applications are developed on such API levels as Google Maps and Yahoo Maps. But many parts of map applications are still constructed by coding because these environments don't support high level of automation. Our MapAppGen that we have designed and implemented enhances the Web-based map application productivity by generating the map related modules that can be applied to the Drupal that is one of popular content management systems(CMS's). Comparing the applications that are constructed by the Drupal-supported GMap or NodeMap, the applications that are constructed by MapAppGen provide information on not only the interested geographical feature but also its related geographical features. MapAppGen uses Google Maps API and Drupal is a module-based system that supports the creation, composition and management of contents. We are now working on automatic generation of mobile map applications with MapAppGen.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.13 no.11
• /
• pp.5471-5478
• /
• 2012

In this paper, we propose a new short-circuit detecting method which can detect bad short-circuits, one of bad types occurring in PCB(Printed Circuit Board), by using SURF(Speeded-Up Robust Features) algorithm. The basic procedure in the proposed method sequentially consists of extracting features from both sample and inputted images by SURF, performing perspective transform by feature matching and matching results, extracting check areas of interest, binary coding and extracting short-circuits, and verifying results. The proposed method focuses on the robustness which can detect bad short-circuits even though the position and angle of PCB are not uniform and arbitrarily placed. Experimental results show that our method enables to detect bad short-circuits regardless of the location and angle of PCB placed variously and validate that the proposed method outperforms the conventional methods detecting bad short-circuits manually on the aspect of both the detection rate and time.