• 전자공학회논문지B
• /
• 제31B권8호
• /
• pp.160-171
• /
• 1994

The wide dynamic range and severely attenuated contrast in mediastinal area appearing in typical chest radiographs have often caused difficulties in effective visualization and diagnosis of lung diseases. This paper proposes a new adaptive image enhancement technique which potentially solves this problem and there by improves observer performance through image processing. In the proposed method image processing is applied to the chest radiograph with different processing parameters for the lung field and mediastinum adaptively since there are much differences in anatomical and imaging properties between these two regions. To achieve this the chest radiograph is divided into the lung and mediastinum by gray level thresholding using the cumulative histogram and the dynamic range compression and local contrast enhancement are carried out selectively in the mediastinal region. Thereafter a gray scale transformation is performed considering the JND(just noticeable difference) characteristic for effective image displa. The processed images showed apparenty improved contrast in mediastinum and maintained moderate brightness in the lung field. No artifact could be observed. In the visibility evaluation experiment with 5 radiologists the processed images with better visibility was observed for the 5 important anatomical structures in the thorax.

• 한국멀티미디어학회논문지
• /
• 제13권5호
• /
• pp.702-709
• /
• 2010

Mass형 종양 분할의 성능은 mass의 초기 위치에 큰 영향을 받는다. 따라서 몇몇의 논문들은 방사선 전문의로부터 획득한 mass의 초기 위치를 이용하여 종양의 분할을 진행하였다. 그러나, 본 논문은 mass 검출을 위한 부가정보 없이 디지털 마모그램만을 이용한 컴퓨터 지원 진단 시스템을 구성하여 방사선 전문의에게 mass로 추정되는 곳의 위치를 제시함을 목표로 한다. 제안된 시스템은 영역 확장 기법과 열림 연산을 통한 유방 영역 분할, 분할된 유방영역에서 mass 특성을 갖는 위치의 시드 설정, 설정된 시드 기반 레벨 셋을 통한 mass 영역 분할로 구성된다. Mass 분할을 위한 시드 설정은 부표본화된 유방영상에 대해 블록기반 분산 정보와 마스킹 정보를 이용하는 Mass Scoring Measure(MSM) system을 통하여 수행되었다. 테스트에 사용된 이미지는 DDSM 데이터베이스를 사용하였으며, 실험 결과 종양검출의 정확도는 4 FP/image에서 78%의 민감도를 나타내었고, 상하방향(CC)과 내외사방향(MLO) 이미지를 동시 고려시 92%의 민감도를 보였다.

• Korean Journal of Radiology
• /
• 제23권4호
• /
• pp.402-412
• /
• 2022

Objective: To evaluate the image quality and lesion detectability of lower-dose CT (LDCT) of the abdomen and pelvis obtained using a deep learning image reconstruction (DLIR) algorithm compared with those of standard-dose CT (SDCT) images. Materials and Methods: This retrospective study included 123 patients (mean age ± standard deviation, 63 ± 11 years; male:female, 70:53) who underwent contrast-enhanced abdominopelvic LDCT between May and August 2020 and had prior SDCT obtained using the same CT scanner within a year. LDCT images were reconstructed with hybrid iterative reconstruction (h-IR) and DLIR at medium and high strengths (DLIR-M and DLIR-H), while SDCT images were reconstructed with h-IR. For quantitative image quality analysis, image noise, signal-to-noise ratio, and contrast-to-noise ratio were measured in the liver, muscle, and aorta. Among the three different LDCT reconstruction algorithms, the one showing the smallest difference in quantitative parameters from those of SDCT images was selected for qualitative image quality analysis and lesion detectability evaluation. For qualitative analysis, overall image quality, image noise, image sharpness, image texture, and lesion conspicuity were graded using a 5-point scale by two radiologists. Observer performance in focal liver lesion detection was evaluated by comparing the jackknife free-response receiver operating characteristic figures-of-merit (FOM). Results: LDCT (35.1% dose reduction compared with SDCT) images obtained using DLIR-M showed similar quantitative measures to those of SDCT with h-IR images. All qualitative parameters of LDCT with DLIR-M images but image texture were similar to or significantly better than those of SDCT with h-IR images. The lesion detectability on LDCT with DLIR-M images was not significantly different from that of SDCT with h-IR images (reader-averaged FOM, 0.887 vs. 0.874, respectively; p = 0.581). Conclusion: Overall image quality and detectability of focal liver lesions is preserved in contrast-enhanced abdominopelvic LDCT obtained with DLIR-M relative to those in SDCT with h-IR.

• 대한의용생체공학회:의공학회지
• /
• 제10권1호
• /
• pp.1-10
• /
• 1989

An automatic lung nodule detection algorithm was applied for digital radiographic images using Bit Slice Processor. In this algorithm, signal enhancing filtering and signal suppressing filtering were performed on the given digital chest image, respectively. Then we grit the dirt- frrence image from these filtered images, and hi-level island images were obtained by applying various threshold values. From the island images, we decided the suspicious nodules using size and circularity test, and marked them to alert radiologists. The performance of the atgorithm was analyzed with respect to the size, contrast and position of digitally synthesized nodules. This method presented 45.8% of true positive ratio for the nodules of lOw in diameter with 12-16 pixel value differnces.

• 대한방사선기술학회지:방사선기술과학
• /
• 제40권4호
• /
• pp.575-580
• /
• 2017

Moire artifact는 미세병변과 주파수 대역이 중첩되기 때문에 moire artifact를 제거하는 Image processing software를 사용할 경우 미세 병변의 손실을 가져올 수 있다. 본 연구에서는 SD(Sprague Dawley) Rat에 microcalcification과 microfracture와 같은 미세병변을 임의로 형성하여 영상화하고, reference 영상과 grid suppression software를 사용한 영상, optimizied grid 영상을 비교 분석하였다. 영상은 두 명의 영상의학과 전문의가 컨소시엄을 형성하여 판독하였고, 판독 결과 값은 McNemar's test 이용하여 평가하였다. 73개의 microcalcifications 중 Grid suppression후 13 cases에서, optimized grid를 사용한 영상은 3 cases에서 영상의 손실이 확인되어 Grid suppression후의 영상이 통계적으로 유의하게 영상 손실을 발생하고 있음을 보여주고 있다(p=0.021). 총 53개의 fracture line은 Grid suppression을 시행한 후 영상에서 19 cases가 영상의 손실을 보였고, optimized grid를 사용한 영상에서는 영상손실이 없는 것으로 판독되었다. 따라서 미세병변을 진단하는 영상에 있어 moire artifact를 제거하기 위한 grid suppression software 사용은 신중하게 고려되어야 할 것이다.

• 한국멀티미디어학회논문지
• /
• 제25권8호
• /
• pp.1233-1241
• /
• 2022

Tumors of the brain are the deadliest, with a life expectancy of only a few years for those with the most advanced forms. Diagnosing a brain tumor is critical to developing a treatment plan to help patients with the disease live longer. A misdiagnosis of brain tumors will lead to incorrect medical treatment, decreasing a patient's chance of survival. Radiologists classify brain tumors via biopsy, which takes a long time. As a result, the doctor will need an automatic classification system to identify brain tumors. Image classification is one application of the deep learning method in computer vision. One of the deep learning's most powerful algorithms is the convolutional neural network (CNN). This paper will introduce a novel deep learning structure and image gradient to classify brain tumors. Meningioma, glioma, and pituitary tumors are the three most popular forms of brain cancer represented in the Figshare dataset, which contains 3,064 T1-weighted brain images from 233 patients. According to the numerical results, our method is more accurate than other approaches.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 추계학술대회
• /
• pp.655-657
• /
• 2021

인공지능 알고리즘을 이용한 유방암의 조기진단에 관련된 연구는 최근들어 활발하게 진행되고 있으나, 사용자의 목적에 맞는 처리속도 및 정확도 등에 다양한 한계점을 보인다. 이러한 문제를 해결하기 위해, 본 논문에서는 ImageNet에서 학습된 ResNet 모델을 현미경 기반 암세포 이미지에서 활용이 가능한 다단계 전이 학습을 제안하고, 이를 다시 전이 학습하여 초음파 유방암 영상을 양성 및 악성으로 분류하는 실험을 진행하였다. 제안된 다단계 전이 학습 알고리즘은 초음파 유방암 영상을 분류하였을 때 96% 이상의 정확도를 보였으며, 향후 암 세포주 및 실시간 영상처리 등의 추가를 통해 보다 높은 활용도와 정확도를 보일 것으로 기대한다.

• Imaging Science in Dentistry
• /
• 제30권4호
• /
• pp.265-274
• /
• 2000

Purpose: The aims of the present study are to determine the optimum compression rate in terms of file size reduction and diagnostic quality of the images after compression and evaluate the transmission speed of original or each compressed image. Materials and Methods: The material consisted of 24 extracted human premolars and molars. The occlusal surfaces and proximal surfaces of the teeth had a clinical disease spectrum that ranged from sound to varying degrees of fissure discoloration and cavitation. The images from Digora system were exported in TIFF and the images from conventional intraoral film were scanned and digitalized in TIFF by Nikon SF-200 scanner (Nikon, Japan). And six compression factors were chosen and applied on the basis of the results from a pilot study. The total number of images to be assessed were 336. Three radiologists assessed the occlusal and proximal surfaces of the teeth with 5-rank scale. Finally diagnosed as either sound or carious lesion by one expert oral pathologist. And sensitivity, specificity and k value for diagnostic agreement was calculated. Also the area (Az) values under the ROC curve were calculated and paired t-test and oneway ANOVA test was performed. Thereafter, transmission time of the image files of the each compression level was compared with that of the original image files. Results: No significant difference was found between original and the corresponding images up to 7% (1 : 14) compression ratio for both the occlusal and proximal caries (p<0.05). JPEG3 (1 : 14) image files are transmitted fast more than 10 times, maintained diagnostic information in image, compared with original image files. Conclusion: 1 : 14 compressed image file may be used instead of the original image and reduce storage needs and transmission time.

• Imaging Science in Dentistry
• /
• 제49권2호
• /
• pp.139-151
• /
• 2019

Purpose: This study was performed to investigate the effect of exposure parameters on image quality obtained using a cone-beam computed tomography (CBCT) scanner and the relationship between physical factors and clinical image quality depending on the diagnostic task. Materials and Methods: CBCT images of a SedentexCT IQ phantom and a real skull phantom were obtained under different combinations of tube voltage and tube current (Alphard 3030 CBCT scanner, 78-90 kVp and 2-8 mA). The images obtained using a SedentexCT IQ phantom were analyzed technically, and the physical factors of image noise, contrast resolution, spatial resolution, and metal artifacts were measured. The images obtained using a real skull phantom were evaluated for each diagnostic task by 6 oral and maxillofacial radiologists, and each setting was classified as acceptable or unacceptable based on those evaluations. A statistical analysis of the relationships of exposure parameters and physical factors with observer scores was conducted. Results: For periapical diagnosis and implant planning, the tube current of the acceptable images was significantly higher than that of the unacceptable images. Image noise, the contrast-to-noise ratio (CNR), the line pair chart on the Z axis, and modulation transfer function (MTF) values showed statistically significant differences between the acceptable and unacceptable image groups. The cut-off values obtained using receiver operating characteristic curves for CNR and MTF 10 were useful for determining acceptability. Conclusion: Tube current had a major influence on clinical image quality. CNR and MTF 10 were useful physical factors that showed significantly associations with clinical image quality.

• Korean Journal of Radiology
• /
• 제22권2호
• /
• pp.281-290
• /
• 2021

Objective: To assess the performance of content-based image retrieval (CBIR) of chest CT for diffuse interstitial lung disease (DILD). Materials and Methods: The database was comprised by 246 pairs of chest CTs (initial and follow-up CTs within two years) from 246 patients with usual interstitial pneumonia (UIP, n = 100), nonspecific interstitial pneumonia (NSIP, n = 101), and cryptogenic organic pneumonia (COP, n = 45). Sixty cases (30-UIP, 20-NSIP, and 10-COP) were selected as the queries. The CBIR retrieved five similar CTs as a query from the database by comparing six image patterns (honeycombing, reticular opacity, emphysema, ground-glass opacity, consolidation and normal lung) of DILD, which were automatically quantified and classified by a convolutional neural network. We assessed the rates of retrieving the same pairs of query CTs, and the number of CTs with the same disease class as query CTs in top 1-5 retrievals. Chest radiologists evaluated the similarity between retrieved CTs and queries using a 5-scale grading system (5-almost identical; 4-same disease; 3-likelihood of same disease is half; 2-likely different; and 1-different disease). Results: The rate of retrieving the same pairs of query CTs in top 1 retrieval was 61.7% (37/60) and in top 1-5 retrievals was 81.7% (49/60). The CBIR retrieved the same pairs of query CTs more in UIP compared to NSIP and COP (p = 0.008 and 0.002). On average, it retrieved 4.17 of five similar CTs from the same disease class. Radiologists rated 71.3% to 73.0% of the retrieved CTs with a similarity score of 4 or 5. Conclusion: The proposed CBIR system showed good performance for retrieving chest CTs showing similar patterns for DILD.