• Title/Summary/Keyword: 부위분류

Search Result 733, Processing Time 0.038 seconds

Hand gesture recognition based on RGB image data (RGB 영상 데이터 기반 손동작 인식)

  • Kim, Gi-Duk
    • Proceedings of the Korean Society of Computer Information Conference
    • /
    • 2021.07a
    • /
    • pp.15-16
    • /
    • 2021
  • 본 논문에서는 RGB 영상 데이터를 입력으로 하여 mediapipe의 손 포즈 추정 알고리즘을 적용해 손가락 관절 및 주요 부위의 위치를 얻고 이를 기반으로 딥러닝 모델에 학습 후 손동작 인식 방법을 제안한다. 연속된 프레임에서 한 손의 손가락 주요 부위 간 좌표를 얻고 차분 벡터의 x, y좌표를 저장한 후 Conv1D, Bidirectional GRU, Transformer를 결합한 딥러닝 모델에 학습 후 손동작 인식 분류를 하였다. IC4You Gesture Dataset 의 한 손 동적 데이터 9개 클래스에 적용한 결과 99.63%의 손동작 인식 정확도를 얻었다.

  • PDF

Classification of Motor Imagery EEG Signals Based on Non-homogeneous Spatial Filter Optimization (비 동질 공간 필터 최적화 기반의 동작 상상 EEG 신호 분류)

  • Kam, Tae-Eui;Lee, Seong-Whan
    • Proceedings of the Korean Information Science Society Conference
    • /
    • 2011.06a
    • /
    • pp.469-472
    • /
    • 2011
  • 신체 부위를 움직이는 상상을 할 때, 일반적으로 뇌의 감각 및 운동 피질 영역에서 특정 주파수 대역의 EEG(Electroencephalography) 신호의 세기가 감소하거나 증가하는 ERD(Event-Related Desynchronization)/ERS(Event-Related Synchronization) 현상이 발생한다. 하지만 ERD/ERS는 현상은 피험자에 의존적이고 매시도마다 큰 차이를 보인다. 이러한 문제를 해결하기 위해, 본 논문에서 각 시간-주파수 공간에 대하여 서로 다른 공간 필터를 구성하는 비 동질(non-homogeneous) 공간 필터 최적화 방법을 제안한다. EEG 신호는 시간에 대하여 비정상적(non-stationary) 특징을 가지기 때문에 제안하는 방법과 같이 시간에 따라 변화하는 ERD/ERS 특징을 반영하여 공간적 특징을 추출하는 방법은 시간에 대한 변화를 고려하지 않은 기존의 방법보다 우수한 성능을 보인다. 본 논문에서는 International BCI Competition IV에서 제공하는 4가지 동작 상상(왼손, 오른손, 발, 혀)에 대한 EEG 신호 데이터를 사용하여 동작 상상 분류 실험을 하고 이 결과를 기존의 타 방법들과 비교 분석하였다. 실험 결과, 피험자에 따라 서로 다른 시간-주파수 특징이 추출됨을 확인하였고, 최적화된 공간 필터들이 시간에 따라 변화하는 것을 확인하였다. 또한 이러한 특징을 이용하여 분류를 수행하였을 때, 더욱 우수한 분류 결과를 보임을 확인하였다.

A Study on the Auto Lumbar Spine Classification Model Based on EfficinetNetV2 (EfficientNetV2기반 자동 요추분류 모델에 관한 연구)

  • Chung-sub Lee;Dong-Wook Lim;Si-Hyeong Noh;Chul Park;Chang-Won Jeong
    • Annual Conference of KIPS
    • /
    • 2023.05a
    • /
    • pp.448-450
    • /
    • 2023
  • 본 논문에서는 복부 CT 의료영상에서 근감소증 진단을 위한 지표로 활용하는 요추 3번 슬라이스를 분류하기 위해서 CNN 기반의 EfficientNetV2를 사용하여 자동분류모델을 개발하였다. 이를 위해 먼저 전체 복부 CT 의료영상에서 Thoracic, L1, L2, L3, L4, L5, Sacral 7개의 슬라이스를 검출하도록 하였다. 자동분류모델의 정확성을 측정하기 위해서 Test 데이터셋을 사용하여 Confusion Matrix 결과를 통해 개발된 모델의 성능을 검증한 결과를 보였다. 본 연구결과는 복부 CT 영상에서 기존 L3 레벨의 특정 단면에서 근육량을 측정하는 것에서 다양한 부위에서 측정할 수 있는 장점을 갖게 된다. 그리고 의료영상기반의 근감소증 진단 연구에 도움을 줄 것으로 기대하고 있다.

The taxonomic consideration of petal and sepal micromorphology in Lycopus L. (Mentheae-Lamiaceae) (쉽사리속(Lycopus속, 박하족-꿀풀과) 화판 및 악편 미세형태의 분류학적 검토)

  • Moon, Hye-Kyoung;Hong, Suk-Pyo
    • Korean Journal of Plant Taxonomy
    • /
    • v.34 no.4
    • /
    • pp.273-285
    • /
    • 2004
  • A comparative micromorphology of petal and sepal of 15 species of Lycopus was undertaken to assess their usefulness in species identification and to evaluate their significance in the taxonomy using scanning electron microscope and stereo microscope. Five types of trichome are found within the genus: unicellular cylindrical trichome, unicellular trichome with papillae, simple multicellular trichome, capitate glandular trichome, and peltate glandular trichome. The types, distribution, and density of the trichomes show considerable variation among the taxa. The distributional pattern of the unicellular cylindrical trichomes on inner side of corolla is differed among the investigated taxa, and can be classified into four patterns. The apex shape of sepal and the incised position of calyx are also various and be recognizedas four patterns. The shape of cell composing in both outer and inner sides of petal and sepal are shown variously. Finally, the taxonomic significance of micromorphology of the petal and sepal in identification and elucidation of the genus Lycopus, especially among the species is also briefly discussed.

Phylogenetic Analysis of the Genus Phellinus by Comparing the Sequences of Internal Transcribed Spacers and 5.8S Ribosomal DNA (Ribosomal DNA의 Internal Transcribed Spacer(ITS) 부위의 염기서열분석에 의한 Phellinus속의 계통분석에 관한 연구)

  • Chung, Ji-Won;Kim, Gi-Young;Ha, Myung-Gui;Lee, Tae-Ho;Lee, Jae-Dong
    • The Korean Journal of Mycology
    • /
    • v.27 no.2 s.89
    • /
    • pp.124-131
    • /
    • 1999
  • This study was carried out to identify the phylogenetic relationship among Phellinus species by comparing the DNA sequences of the 5.8S ribosomal DNA (rDNA) and the internal transcribed spacers (ITSs), ITS1 and ITS2 regions. Two primers from the 3' end of 18S rDNA and the 5' end of 28S rDNA sequences were chosen to amplify the specific ITS regions of Phellinus spp. Phellinus strains used in the study were divided into four clusters by the phylogenetic tree based on the amplified regions of ITS and 5.8S rDNA sequences. The first cluster consist of Phellinus hartigii IMSNU 32041 and Phellinus robustus IMSNU 32068, and the second cluster consists of Phellinus linteus strains and Phellinus weirianus IMSNU 32021. Phellinus laevigatus KCTC 6229, KCTC 6230 and Phellinus igniarius KCTC 6227, KCTC 6228 belong to the third cluster. Finally, Phellinus chrysoloma KCTC 6225 and Phellinus chrysoloma KCTC 6226 are the fourth cluster. In the second cluster the differentiation between Phellinus linteus strains and Phellinus weirianus species were not possible by the comparison of the ITS sequences. These results revealed that Phellinus linteus and Phellinus weirianus cannot be established the concept of species level only by the ITS sequences. Therefore, both physiological and molecular biological methods as well as the sequences of type strains are necessary to classify the strains of these two species accurately. The comparison of the ITS sequences of four Phellinus species indicated that the sequences of the ITS1 generally are more divergent than those of the ITS2. Although the ITS sequences are varied in some species, the conserved regions in both ITS1 and ITS2 are useful tool to differentiate the species. Phellinus linteus and related species have their specific sequences in the ITS1 compared to the other species.

  • PDF

The Study of CYFRA 21-1 and Epidermal Growth Factor Receptor Levels in Cancer Tissue of Bronchogenic Carcinoma Patients (폐암환자의 암조직내 CYFRA 21-1과 Epidermal Growth Factor Receptor의 측정치에 대한 연구)

  • 김대연;김송명
    • Journal of Chest Surgery
    • /
    • v.30 no.9
    • /
    • pp.854-861
    • /
    • 1997
  • CYPRA 21-1 is known to be a cytokeratin 19 fragment, and it can be detected by using two specific monoclonal antibodies (KS 19-1 and BM 19-21) and can be clinically applied as a useful circulating tumor marker The epidermal growth factor receptor (EGF-R) expression was evaluated and characterized by its tyrosine protein kinase activity and by its ligand-stimulated autophosphorylation, a property shared with other peptide growth factor receptors. Autocrine or para'urine action was initiated by a growth factor, or by a transforming growth factor o, which had an extensive homology with EGP and which also stimulated tyrosine kinase activity on the EGF-R. The CYFRA 21-1 and the EGF-R levels in 30 patients with primary lung tumors were investigated. There were 24 patients with squamous cell carcinomas and 6 patients with adenocarcinomas. Specimen 5 mm3 in size were sampled at three different locations ; the main lesion, the boundary between the lesion and the unaffected tissue, and the unaffected tissue of the patients. The results were as follows 1. The CYPRA 21-1 concentration in the cancer boundary, the most malignant region,(348.6 : 89.9 ng/ml) was the lowest value. The CYFRA 21-1 concentration in unaffected tissue,(718.4$\pm$77.8 ng/ml) was higher than that in the main lesion. which had intact cellularity. 2. The EGF-R concentration in the main lesion was higher than that in the unaffected tissue, and the EGF-R concentration in a squamous cell cacinoma was higher than that in an adenocarcinoma. also, the EGF-R concentration in the cancer b undary was highest at stage 1, ll. The EGF-R concentration was higher in the main cancer lesion that in the unaffected tissue at stage 111, IV. 3. The CYFRA 21-1 was a cytoplasmic skeleton and the EGF-R was a cell-wall component; there was no correlation. In conclusion, CYFRA 21-1 was abundant in the cytoplasm but had a higher concentration in the unaffected tissue than in the main cancer lesion. The CYFRA 21-1 concentration of the tissue did not reflect the amount of cancer activity, the EGP-R was located in the cell membrane, the level of tissue that reflects cancer activity, so the main cancer lesion had a higher concentration than the unaffected tissue. CYFRA 21-1 is not a useful tumor maker at the tissue level. Because the EGF-R concentration re(looted the cancer activity, its a useful tumor marker for lung cancer.

  • PDF

Germ Cell Differentiations during Spermatogenesis and Ultrastructural Characteristics of Mature Sperms in Male Protothaca (Notochione) jedoensis (Bivalvia: Veneridae) (수컷 살조개, Protothaca (Notochione) jedoensis (Bivalvia: Veneridae)의 정자형성과정 중 생식세포 분화와 성숙정자의 미세구조적 특징)

  • Kim, Jin-Hee;Park, Young-Jae;Lee, Ki-Young;Choi, Moon-Sul;Seo, Won-Jae;Chung, Ee-Yung
    • Development and Reproduction
    • /
    • v.14 no.4
    • /
    • pp.269-279
    • /
    • 2010
  • Some characteristics of germ cell differntiations and the function of accessory cells during spermatogenesis, and mature sperm ultrastructure in male Protothaca (N.) jedoensis were investigated by transmission electron microscope observations. The morphology of the spermatozoa of this species has a primitive type and is similar to those of other species in the subclass Heterodonta. Accessory cells, which are connected to adjacent germ cells, are involved in the supplying of the nutrients for germ cell development. The morphologies of the sperm nucleus and the acrosome of this species are the cylindrical type and cap shape, respectively. Spermatozoa are approximately $46{\sim}50{\mu}m$ in length including a long sperm nucleus (about $2.44{\mu}m$ in length), an acrosome (about $0.45{\mu}m$ in length), and tail flagellum (about $42{\sim}46{\mu}m$). The axoneme of the sperm tail shows a 9+2 structure. As some characteristics of the acrosomal vesicle structures, the basal and lateral parts of basal rings show electron opaque part (region), while the anterior apex part of the acrosomal vesicle shows electron lucent part (region). These characteristics of the acrosomal vesicle were found in the family Veneridae and other several families in the subclass Heterodonta. These common characteristics of the acrosomal vesicle in the subclass Heterodonta can be used for phylogenetic and systematic analysis as a taxonomic key or a significant tool. The number of mitochondria in the midpiece of the sperm of this species are four, as one of common characteristics appear in most species in the family Veneridae and other families in the subclass Heterodonta. However, exceptionally, only three species in Veneridae of the subclass Heterodonta contain 5 mitochondria. The number of mitochondria in the sperm midpiece can be used for the taxonomic analysis of the family or superfamily levels as a systematic key or an important tool.

Prevalence of Emergency Traumatic Injuries in 100 Wild Boar Hunting Dogs (100마리의 야생 멧돼지 사냥견에서 응급 외상성 질환의 분류)

  • Ko, Jae-Jin;Kim, Se-Hoon;Ji, Joong-Ryong;Shim, Kwan-Seob;Kim, Nam-Soo
    • Journal of Veterinary Clinics
    • /
    • v.27 no.6
    • /
    • pp.718-722
    • /
    • 2010
  • To investigate patterns and severity of hunting-related emergency traumatic injuries in wild boar hunting dogs. One hundred wild boar hunting dogs with emergency traumatic injuries sustained during wild boar hunting as a result of wild boar attack and accidental shooting of ensnarement in a trap. The retrospective study involved 100 dogs brought to CAMC for treatment of emergency traumatic injury sustained during wild boar hunting in Jeon-buk province from August 2007 to April 2008. Medical information obtained from the medical records included signalment; cause of injury; number, location and severity of injuries; and mortality. The 100 patients displayed 136 injuries (single injury in 71 dogs and multiple injuries in 29 dogs). Causes of the emergency traumatic injuries were wild boar attack (n = 92), accidental shooting (n = 7), and entrapment (n = 1). The thoracic area was the most common site of injury. The most common injury severity score (ISS) was code 2. The mortality rate was 9%, and all deaths involved thoracic injury. Emergency traumatic injuries sustained during wild boar hunting are most commonly thoracic injuries caused by prey attack. The nature of the injuries can differ from those typically encountered by small animal veterinarians.

Comparative seed morphology of Solanaceae in Korea (한국산 가지과의 종자 비교형태)

  • Kong, Min-Jung;Lee, Jong-Su;Hong, Suk-Pyo
    • Korean Journal of Plant Taxonomy
    • /
    • v.41 no.1
    • /
    • pp.35-46
    • /
    • 2011
  • The seed morphology of 22 taxa (12 genera) of Solanaceae in Korea was studied by stereo and scanning electron microscopy, and a detailed description of seed morphological features for all examined taxa is provided. The color of the seed is yellow to black, and the shape is reniform to circular. Its size is 0.56-4.29 mm in length and 0.38-3.20 mm in width. The largest sample was found in Datural metel, while the smallest was in Petunia hybrida. Three different surface types (verrucate, verrucate-reticulate, and reticulate) were recognized in the studied taxa. The top of the anticlinal wall produces appendages known as 'fibrils'. This projection of anticlinal wall was found to be specific to the genus Solanum (S. japonense, S. lyratum, and S. lycopersicum). The morphological features (seed shape, size, color, hilum shape, surface type, and fibrils) of the family Solanaceae in Korea are described and their systematic implications are briefly discussed.

Classification of Brain Magnetic Resonance Images using 2 Level Decision Tree Learning (2 단계 결정트리 학습을 이용한 뇌 자기공명영상 분류)

  • Kim, Hyung-Il;Kim, Yong-Uk
    • Journal of KIISE:Software and Applications
    • /
    • v.34 no.1
    • /
    • pp.18-29
    • /
    • 2007
  • In this paper we present a system that classifies brain MR images by using 2 level decision tree learning. There are two kinds of information that can be obtained from images. One is the low-level features such as size, color, texture, and contour that can be acquired directly from the raw images, and the other is the high-level features such as existence of certain object, spatial relations between different parts that must be obtained through the interpretation of segmented images. Learning and classification should be performed based on the high-level features to classify images according to their semantic meaning. The proposed system applies decision tree learning to each level separately, and the high-level features are synthesized from the results of low-level classification. The experimental results with a set of brain MR images with tumor are discussed. Several experimental results that show the effectiveness of the proposed system are also presented.