• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.313-315
• /
• 2021

Timely and accurate diagnosis of lung diseases using Chest X-ray images has been gained much attention from the computer vision and medical imaging communities. Although previous studies have presented the capability of deep convolutional neural networks by achieving competitive binary classification results, their models were seemingly unreliable to effectively distinguish multiple disease groups using a large number of x-ray images. In this paper, we aim to build an advanced approach, so-called Ensemble Knowledge Distillation (EKD), to significantly boost the classification accuracies, compared to traditional KD methods by distilling knowledge from a cumbersome teacher model into an ensemble of lightweight student models with parallel branches trained with ground truth labels. Therefore, learning features at different branches of the student models could enable the network to learn diverse patterns and improve the qualify of final predictions through an ensemble learning solution. Although we observed that experiments on the well-established ChestX-ray14 dataset showed the classification improvements of traditional KD compared to the base transfer learning approach, the EKD performance would be expected to potentially enhance classification accuracy and model generalization, especially in situations of the imbalanced dataset and the interdependency of 14 weakly annotated thorax diseases.

• Journal of the Korean Society of Radiology
• /
• v.14 no.1
• /
• pp.39-44
• /
• 2020

The purpose of this study was learning about chest X-ray image classification and accuracy research through Deep Learning using big data technology with Convolution Neural Network. Normal 1,583 and Pneumonia 4,289 were used in chest X-ray images. The data were classified as train (88.8%), validation (0.2%) and test (11%). Constructed as Convolution Layer, Max pooling layer size 2×2, Flatten layer, and Image Data Generator. The number of filters, filter size, drop out, epoch, batch size, and loss function values were set when the Convolution layer were 3 and 4 respectively. The test data verification results showed that the predicted accuracy was 94.67% when the number of filters was 64-128-128-128, filter size 3×3, drop out 0.25, epoch 5, batch size 15, and loss function RMSprop was 4. In this study, the classification of chest X-ray Normal and Pneumonia was predictable with high accuracy, and it is believed to be of great help not only to chest X-ray images but also to other medical images.

• Journal of Chest Surgery
• /
• v.33 no.6
• /
• pp.521-524
• /
• 2000

The benign teratoma is usually slow growing tumor, but we expirienced a case of primary huge mediastinal benign teratoma that had grown very rapidly, maximally during 3 years. The 14-year-old female patient was admitted to our hospital because of abnormal chest X-ray that showed 10$\times$10cm sized well definded mass with multiple calcificactions. but the mass was not present in chest X-ray perfomed on 3 years prior to admission. Under the diagnosis of teratoma, complete surgical resection was done by the left thoracotomy. The result of pathology was benign teratoma.

• Journal of the Korean Society of Radiology
• /
• v.14 no.6
• /
• pp.773-780
• /
• 2020

With the development of the 4th industrial, research is being conducted to prevent diseases and reduce damage in various fields of science and technology such as medicine, health, and bio. As a result, artificial intelligence technology has been introduced and researched for image analysis of radiological examinations. In this paper, we will directly apply a deep learning model for classification and detection of pneumonia using chest X-ray images, and evaluate whether the deep learning model of the Inception series is a useful model for detecting pneumonia. As the experimental material, a chest X-ray image data set provided and shared free of charge by Kaggle was used, and out of the total 3,470 chest X-ray image data, it was classified into 1,870 training data sets, 1,100 validation data sets, and 500 test data sets. I did. As a result of the experiment, the result of metric evaluation of the Inception V3 deep learning model was 94.80% for accuracy, 97.24% for precision, 94.00% for recall, and 95.59 for F1 score. In addition, the accuracy of the final epoch for Inception V3 deep learning modeling was 94.91% for learning modeling and 89.68% for verification modeling for pneumonia detection and classification of chest X-ray images. For the evaluation of the loss function value, the learning modeling was 1.127% and the validation modeling was 4.603%. As a result, it was evaluated that the Inception V3 deep learning model is a very excellent deep learning model in extracting and classifying features of chest image data, and its learning state is also very good. As a result of matrix accuracy evaluation for test modeling, the accuracy of 96% for normal chest X-ray image data and 97% for pneumonia chest X-ray image data was proven. The deep learning model of the Inception series is considered to be a useful deep learning model for classification of chest diseases, and it is expected that it can also play an auxiliary role of human resources, so it is considered that it will be a solution to the problem of insufficient medical personnel. In the future, this study is expected to be presented as basic data for similar studies in the case of similar studies on the diagnosis of pneumonia using deep learning.

• Journal of Chest Surgery
• /
• v.40 no.5 s.274
• /
• pp.392-394
• /
• 2007

A 14-year-old male patient was admitted for an abnormal chest X-ray. A chest computed tomogram showed a cystic mass in the anterior mediastinum and spleen, $14\times14cm$ and $2\times2cm$ in size respectively. Complete removal of the mediastinal lesion was achieved by a median sternotomy. The final histologic diagnosis of the lesion was cystic lymphangioma. There was no evidence of tumor recurrence until a postoperative period of 14 months.

• Journal of radiological science and technology
• /
• v.43 no.2
• /
• pp.65-70
• /
• 2020

PA(postero-anterior) and AP(antero-posterior) chest projections are the most sought-after types of all kinds of projections. But if a radiological technologist puts wrong information about the position in the computer, the orientation of left and right side of an image would be reversed. In order to solve this problem, we utilized CNN(convolutional neural network) which has recently utilized a lot for studies of medical imaging technology and rule-based system. 70% of 111,622 chest images were used for training, 20% of them were used for testing and 10% of them were used for validation set in the CNN experiment. The same amount of images which were used for testing in the CNN experiment were used in rule-based system. Python 3.7 version and Tensorflow r1.14 were utilized for data environment. As a result, rule-based system had 66% accuracy on evaluating whether the orientation reversal on chest x-ray image. But the CNN had 97.9% accuracy on that. Being overcome limitations by CNN which had been shown on rule-based system and shown the high accuracy can be considered as a meaningful result. If some problems which can occur for tasks of the radiological technologist can be separated by utilizing CNN, It can contribute a lot to optimize workflow.

• Journal of radiological science and technology
• /
• v.22 no.1
• /
• pp.49-53
• /
• 1999

To inquire its usefulness of the clinical application of intelligent replenishment system of automatic X-ray film processor based on film density, we processed the serial 300 sheets of radiographic film of chest [$14{\times}14"$, HR-C type] and bone [elbow & ankle($8{\times}10"$), skull($10{\times}12"$), hand & foot($11{\times}14"$), pelvis($14{\times}17"$), HR-G type, 68, 70, 77, 85 sheets respectively]. We analyzed the characteristic corves, relative speeds, average gradients and base plus fog densities every twenty five sheets. We also evaluated the developer and fixer replenishment volumes every that time. In the chest and bone radiograph two all, the characteristic curves were little change, and the relative speeds, average gradients and base plus fog densities were within the maximum control limits. The average developer replenishment volumes were about 43m1/sheet and 39m1/sheet respectively. It brings decreased results about 29% in comparison with the conventional replenishment system. In our experiences, we conclude that the intelligent replenishment system of automatic X-ray film processor based on film density maintains image quality consistently, decreases also the replenishment volumes. Therefore, this system will be resulted in economic and environmental effects, and solve problems of over and low replenishment volume.

• Journal of radiological science and technology
• /
• v.20 no.2
• /
• pp.49-55
• /
• 1997

This study was carried out to investigate radiographical and operating conditions of X-ray units and exposure doses to patients during chest radiography, so that the results could provide basic data used for reducing the exposure dose and for providing the diagnostic information with better quality. The conditions and exposure doses of 100 X-ray units mainly used for chest radiography were examined and also 100 radiological technologists mainly handling those apparatus at 76 medical facilities in Pusan were surveyed using a questionnaire from October 1 to December 31 in 1995. The following results were obtained from the study : 1. It was found that most units were capable of taking a high tube voltage radiography by showing 67% of the units equipped with the maximum tube voltage of 150 kV, 94% with more than 500 mA for the rating capacity and 85% with the full wave type of a signal phase. 2. For actual chest radiographical conditions, however, 80% of the units were operated at $60{\sim}100\;kVp$ and only 14% at 100 kVp and over for the high tube voltage. 3. The average exposure time was less than 0.1 second, and eighty four percent of the units adapted the X-ray tube currents ranging from 200 to 300 mA, 80% the focus-film distances between 180 and 210 cm, and 63% the focus sizes of more than 2.0 mm. 4. Most units(98%) employed additional filters made of aluminum, 75% the thickness of filters less than 2.0 mm, and only 2 units the compound filters. 5. Ortho chromatic system was only adopted in 13% of screen film system for the units, and 73% used the grid ratio at 8 : 1 for the low tube voltage during chest radiography. 6. The average exposure dose of all X-ray units during chest radiography was $371\;{\mu}Sv$ with a difference of about 16 times between the minimum to the maximum, and $386\;{\mu}Sv$ both at hospitals and at health centers, followed by $380\;{\mu}Sv$ at general hospitals and $263\;{\mu}Sv$ at university hospitals without showing any statistically significant differences. In conclusion, since patients during chest radiography at medical facilities in Pusan exposed to high levels of radiation, it is recommended that appropriate added filters and grids necessary for the high tube voltage radiography and high-speed screen systems should be adopted and used as soon as possible in order to reduce exposure dose to the patients.

• Journal of Pharmacopuncture
• /
• v.4 no.3
• /
• pp.59-67
• /
• 2001

Objective: The purpose of this report is to prove the clinical effect of Platycodon grandiflorum aqueous extract on pneumoniae patients. Methods: We used the aqueous extract of Platycodon grandiflorum to treat two pneumoniae patients. It was injected into five acupuncture points, which was Chondol(天突:CV22) 1 point, Pyesu(肺兪 : BL13) 2 point, and Kworumsu(厥陰兪: BL14) 2 point. Results & conclusions: We have used the aqueous extract of 24-year-old JK for treating the patients suffering from lung diseases, and have experienced the actual effects. Of the treated, two pneumonia-involved patients showed apparent improvement in simple chest X-ray and clinical symptoms. The patients were treated with JK (Jang-saeng platycodon) aqueous extract 25 and 22 times individually. The results were as follows. 1. The symptoms including coughing, phlegm, and fever were improved in two cases. 2. The lung infiltration in simple chest X-ray decreased and the WBC count was kept within normal range in two cases. 3. Side effect such as itching was not found in the process of JK aqueous extract treatment. 4. The criteria for pneumonia are fever, coughing with purulent phlegm, pleural chest pain, the evidence of new infiltration in simple chest X-ray, sign of lung sclerosis in auscultation, increase of WBC count, etc. But they may not be the proper objective diagnostic standards. So we had trouble in statistic process and numerical interpretation. Putting these results together, the JK aqueous extract is considered to be effective in treating patients for pneumonia, and the continuous research and accumulation of data is needed.

• Journal of Chest Surgery
• /
• v.26 no.12
• /
• pp.959-961
• /
• 1993

Localized fibrous tumor of pleura is submesothelial origin and related terms with localized mesothelioma, giant sarcoma of visceral pleura, post-inflammatory tumor of the pleura, pleural fibroma, submesothelial fibroma. This tumor is rare. We experienced a case of localized fibrous tumor.This 66 years old female was admitted with 2 years left persistant flank pain and mild dyspnea. Chest X-ray and CT scan showed a 12x10cm well-defined huge mass in the left subpulmonic area, and not metastatic lesion of any organs.Exploratory thoracotomy was done and a 14x10x8cm [650gm weight] sized mass was excised.The patient was discharged without any complications postoperatively.