• International Journal of Computer Science & Network Security
• /
• 제21권6호
• /
• pp.1-6
• /
• 2021

Recently, pre-trained convolutional neural network CNNs have been widely used and applied for medical image classification. These models can utilised in three different ways, for feature extraction, to use the architecture of the pre-trained model and to train some layers while freezing others. In this study, the ResNet18 pre-trained CNNs model is used for feature extraction, followed by the support vector machine for multiple classes to classify medical images from multi-classes, which is used as the main classifier. Our proposed classification method was implemented on Kvasir and PH2 medical image datasets. The overall accuracy was 93.38% and 91.67% for Kvasir and PH2 datasets, respectively. The classification results and performance of our proposed method outperformed some of the related similar methods in this area of study.

• 한국시뮬레이션학회논문지
• /
• 제18권4호
• /
• pp.197-206
• /
• 2009

의료정보 시스템은 의료영상과 진단정보를 공유할 수 있는 환경을 제공해주는 효과적인 진단 보조 도구이지만 단순히 정보의 저장과 전송만을 제공한다. 이러한 단점을 해결하고 진단활동의 효율성을 높이기 위해서는 의료영상 분류 및 검색 시스템이 필요하다. 의료영상 분류 및 검색 시스템은 질환 영상과 유사한 영상을 제공함으로써 진단활동의 효율성을 높이고, 다양한 사례 확인을 통하여 보다 전문적인 의료활동을 제공할 수 있다. 그러나 기존의 영상 분류 및 검색 시스템은 영상의 표면적인 정보만을 이용하므로 영상이 내포하는 의미를 파악하기 어렵다. 그러므로 영상의 표면적인 정보뿐만 아니라 영상을 구성하는 요소들의 관계를 파악하여 영상을 분류할 수 있는 의료영상 분류 시스템이 필요하다. 본 논문에서 제안한 기법은 뇌 자기공명영상에서 영상의 표면적인 정보와 공간정보를 추출하여 뇌 자기공명영상을 학습하고 분류한다. 영상의 표면적인 정보는 영상 자체가 갖는 색상, 모양 등의 정보로 하위 영상정보라 하고, 영상의 논리정보를 상위 영상정보라 한다. 본 논문에서는 하위 영상정보와 상위 영상정보를 추출할 때 뇌의 해부학적 명칭과 구조를 활용하였다. 하위 영상정보는 뇌 영상의 부분 영역들에 대한 해부학적 명칭을 부여하기 위해 활용되고, 상위 영상정보는 명칭이 부여된 부분 영역들의 관계를 활용하여 정보를 추출한다. 각 정보는 학습과 분류에 사용된다. 실험에서는 질환을 갖는 뇌 자기공명영상을 활용하였다.

• Imaging Science in Dentistry
• /
• 제32권3호
• /
• pp.129-134
• /
• 2002

Purpose : To implement the database system of oral and maxillofacial radiological images using a commercial medical image management software with personally developed classification code. Materials and methods : The image database was built using a slightly modified commercial medical image management software, Dr. Image v.2.1 (Bit Computer Co., Korea). The function of wild card '*' was added to the search function of this program. Diagnosis classification codes were written as the number at the first three digits, and radiographic technique classification codes as the alphabet right after the diagnosis code. 449 radiological films of 218 cases from January, 2000 to December, 2000, which had been specially stored for the demonstration and education at Dept. of OMF Radiology of Dankook University Dental Hospital, were scanned with each patient information. Results: Cases could be efficiently accessed and analyzed by using the classification code. Search and statistics results were easily obtained according to sex, age, disease diagnosis and radiographic technique. Conclusion : Efficient image management was possible with this image database system. Application of this system to other departments or personal image management can be made possible by utilizing the appropriate classification code system.

• 한국컴퓨터정보학회논문지
• /
• 제23권4호
• /
• pp.147-153
• /
• 2018

In this paper, we propose an efficient WBC 14-Diff classification which performs using the WBC-ResNet-152, a type of CNN model. The main point of view is to use Super-pixel for the segmentation of the image of WBC, and to use ResNet for the classification of WBC. A total of 136,164 blood image samples (224x224) were grouped for image segmentation, training, training verification, and final test performance analysis. Image segmentation using super-pixels have different number of images for each classes, so weighted average was applied and therefore image segmentation error was low at 7.23%. Using the training data-set for training 50 times, and using soft-max classifier, TPR average of 80.3% for the training set of 8,827 images was achieved. Based on this, using verification data-set of 21,437 images, 14-Diff classification TPR average of normal WBCs were at 93.4% and TPR average of abnormal WBCs were at 83.3%. The result and methodology of this research demonstrates the usefulness of artificial intelligence technology in the blood cell image classification field. WBC-ResNet-152 based morphology approach is shown to be meaningful and worthwhile method. And based on stored medical data, in-depth diagnosis and early detection of curable diseases is expected to improve the quality of treatment.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2007년도 추계종합학술대회
• /
• pp.301-304
• /
• 2007

A realization for image searching engine with moving objects identification and classification is presented in this paper. The identification algorithm is applied to extract difference image between input image and the reference image, and the classification is used the region segmentation. That is made the database for the searching engine. The experimental result of the realized system enables to search for human and animal at time intervals to use a surveillant system at inside environment.

• 디지털산업정보학회논문지
• /
• 제16권1호
• /
• pp.55-65
• /
• 2020

With the development of display hardware, image interpolation techniques have been used in various fields such as image zooming and medical imaging. Traditional image interpolation methods, such as bi-linear interpolation, bi-cubic interpolation and edge direction-based interpolation, perform interpolation in the spatial domain. Recently, interpolation techniques in the discrete cosine transform or wavelet domain are also proposed. Using these various existing interpolation methods and machine learning, we propose decision tree classification-based image interpolation methods. In other words, this paper is about the method of adaptively applying various existing interpolation methods, not the interpolation method itself. To obtain the decision model, we used Weka's J48 library with the C4.5 decision tree algorithm. The proposed method first constructs attribute set and select classes that means interpolation methods for classification model. And after training, interpolation is performed using different interpolation methods according to attributes characteristics. Simulation results show that the proposed method yields reasonable performance.

• 대한의용생체공학회:의공학회지
• /
• 제42권6호
• /
• pp.277-286
• /
• 2021

The global market size of AI based SaMD for medical image in 2023 will be anticipated to reach around 620 billion won (518 million dollars). In order for Korean manufacturers to efficiently obtain CE marking for marketing in the EU countries, the paper is to introduce the recommendation and suggestion of how to reclassify SaMD based on classification rules of MDR because, after introducing the Regulation EU 2017/745, classification rules are quite modified and newly added compared to the Directive 93/42/EEC. In addition, the paper is to provide several rules of MDR that may be applicable to decide the classification of SaMD. Lastly, the paper is to examine and demonstrate various secondary data supported by qualitative data because the paper focuses on the suggestion and recommendation with a public trust on the basis of various secondary data conducted by the analysis of field data. In conclusion, the paper found that the previous classification of SaMD followed by the rule of MDD should be reclassified based on the Regulation EU 2017/745. Therefore, the suggestion and recommendation are useful for Korean manufacturers to comprehend the classification of SaMD for marketing in the EU countries.

• 대한의용생체공학회:의공학회지
• /
• 제39권5호
• /
• pp.213-219
• /
• 2018

Stomach cancer is the most diagnosed cancer in Korea. When gastric cancer is detected early, the 5-year survival rate is as high as 90%. Gastroscopy is a very useful method for early diagnosis. But the false negative rate of gastric cancer in the gastroscopy was 4.6~25.8% due to the subjective judgment of the physician. Recently, the image classification performance of the image recognition field has been advanced by the convolutional neural network. Convolutional neural networks perform well when diverse and sufficient amounts of data are supported. However, medical data is not easy to access and it is difficult to gather enough high-quality data that includes expert annotations. So This paper evaluates the efficacy of transfer learning in gastroscopy classification and diagnosis. We obtained 787 endoscopic images of gastric endoscopy at Gil Medical Center, Gachon University. The number of normal images was 200, and the number of abnormal images was 587. The image size was reconstructed and normalized. In the case of the ResNet50 structure, the classification accuracy before and after applying the transfer learning was improved from 0.9 to 0.947, and the AUC was also improved from 0.94 to 0.98. In the case of the InceptionV3 structure, the classification accuracy before and after applying the transfer learning was improved from 0.862 to 0.924, and the AUC was also improved from 0.89 to 0.97. In the case of the VGG16 structure, the classification accuracy before and after applying the transfer learning was improved from 0.87 to 0.938, and the AUC was also improved from 0.89 to 0.98. The difference in the performance of the CNN model before and after transfer learning was statistically significant when confirmed by T-test (p < 0.05). As a result, transfer learning is judged to be an effective method of medical data that is difficult to collect good quality data.

• 대한의용생체공학회:의공학회지
• /
• 제43권1호
• /
• pp.45-51
• /
• 2022

The purpose of this study is to improve the classification accuracy compared to the existing InceptionV3 model by proposing a new model modified with the fully connected hierarchical structure of InceptionV3, which showed excellent performance in medical image classification. The data used for model training were trained after data augmentation on a total of 1026 chest X-ray images of patients diagnosed with normal heart and Cardiomegaly at Kyungpook National University Hospital. As a result of the experiment, the learning classification accuracy and loss of the InceptionV3 model were 99.57% and 1.42, and the accuracy and loss of the proposed model were 99.81% and 0.92. As a result of the classification performance evaluation for precision, recall, and F1 score of Inception V3, the precision of the normal heart was 78%, the recall rate was 100%, and the F1 score was 88. The classification accuracy for Cardiomegaly was 100%, the recall rate was 78%, and the F1 score was 88. On the other hand, in the case of the proposed model, the accuracy for a normal heart was 100%, the recall rate was 92%, and the F1 score was 96. The classification accuracy for Cardiomegaly was 95%, the recall rate was 100%, and the F1 score was 97. If the chest X-ray image for normal heart and Cardiomegaly can be classified using the model proposed based on the study results, better classification will be possible and the reliability of classification performance will gradually increase.

• 핵의학기술
• /
• 제26권2호
• /
• pp.37-42
• /
• 2022

Purpose Since 1985, the Korean society of nuclear medicine technology (KSNMT) has been engaged in academic activities related to nuclear medicine imaging. From 2017 to 2021, the papers published in the journal were classified by the specific fields to examine the trends in the research and the direction of nuclear medicine in comparison with the papers submitted to the Korean Society of Nuclear Medicine (KSNM) during the same period. Materials and Methods From 2017 to 2021, papers submitted to KSNMT and KSNM were classified and databaseization using the Excel program by submission type, examination equipment, and examination field. Through this data, the number of papers published in journals by year, the number of papers submitted by detailed fields, and key words by era were analyzed and compared. Results The papers included by journal was 57 KSNMT and 280 KSNM. The major large classification of equipment, PET, Planar and SPECT was 26.3%, 21.1%, 19.3% in the KSNMT, KSNM was 49.6%, 6.4%, and 9.3%, with 66.7% and 65.3%, respectively. the major medium classification of equipment, industrial safety, urogenital system, nervous system, and quality control accounted for 54.4% of the total papers of the total ratio in the KSNMT, while the medium classification of oncology, endocrine system, urogenital system, therapy, and nervous system accounted for 61.1% of KSNM. In the major small classification of image acquisition, improvement effect, and exposure management accounted for 70.2% in KSNMT, while the items of image acquisition, report, and improvement effect accounted for 60.7% in KSNM. The major keywords except for equipment-related keywords such as PET/CT, PET/MR, and SPECT were SUV, Planar Image, and Respiration Gating Method in KSNMT and Ga68, Thyroid, and Lymphoma in the KSNM. Conclusion When checking the last 5 years of submissions, we can see that KSNMT is mainly concerned with image acquisition using existing radiotracers, while KSNM has focused on new radiotracers such as ⁶⁸Ga, ¹⁷⁷Lu, etc., and new medical technologies of theranostic. It has been confirmed that more PET-related papers than other examination equipment will account for a greater number of papers, and it is believed that future submissions will also account for a higher proportion of PET-related papers than other equipment.