• Journal of Biomedical Engineering Research
• /
• v.44 no.3
• /
• pp.211-217
• /
• 2023

Recently, many studies have introduced artificial intelligence systems in the surgical process to reduce the incidence and mortality of complications in patients. Bleeding is a major cause of operative mortality and complications. However, there have been few studies conducted on detecting bleeding in surgical videos. To advance the development of deep learning models for detecting intraoperative hemorrhage, three models have been trained and compared; such as, YOLOv5, RetinaNet50, and RetinaNet101. We collected 1,016 bleeding images extracted from five surgical videos. The ground truths were labeled based on agreement from two specialists. To train and evaluate models, we divided the datasets into training data, validation data, and test data. For training, 812 images (80%) were selected from the dataset. Another 102 images (10%) were used for evaluation and the remaining 102 images (10%) were used as the evaluation data. The three main metrics used to evaluate performance are precision, recall, and false positive per image (FPPI). Based on the evaluation metrics, RetinaNet101 achieved the best detection results out of the three models (Precision rate of 0.99±0.01, Recall rate of 0.93±0.02, and FPPI of 0.01±0.01). The information on the bleeding detected in surgical videos can be quickly transmitted to the operating room, improving patient outcomes.

• Genomics & Informatics
• /
• v.21 no.1
• /
• pp.8.1-8.19
• /
• 2023

Cancer of the stomach is the second most frequent cancer-related death worldwide. The survival rate of patients with gastric cancer (GC) remains fragile. There is a requirement to discover biomarkers for prognosis approaches. Helicobacter pylori in the stomach is closely associated with the progression of GC. We identified the genes associated with poor/favorable prognosis in H. pylori-induced GC. Multivariate statistical analysis was applied on the Gene Expression Omnibus (GEO) dataset GSE54397 to identify differentially expressed miRNAs (DEMs) in gastric tissues with H. pylori-induced cancer compared with the H. pylori-positive with non-cancerous tissue. A protein interaction map (PIM) was built and subjected to DEMs targets. The enriched pathways and biological processes within the PIM were identified based on substantial clusters. Thereafter, the most critical genes in the PIM were illustrated, and their prognostic impact in GC was investigated. Considering p-value less than 0.01 and |Log2 fold change| as >1, five microRNAs demonstrated significant changes among the two groups. Gene functional analysis revealed that the ubiquitination system, neddylation pathway, and ciliary process are primarily involved in H. pylori-induced GC. Survival analysis illustrated that the overexpression of DOCK4, GNAS, CTGF, TGF-b1, ESR1, SELE, TIMP3, SMARCE1, and TXNIP was associated with poor prognosis, while increased MRPS5 expression was related to a favorable prognosis in GC patients. DOCK4, GNAS, CTGF, TGF-b1, ESR1, SELE, TIMP3, SMARCE1, TXNIP, and MRPS5 may be considered prognostic biomarkers for H. pylori-induced GC. However, experimental validation is necessary in the future.

• Journal of Biomedical Engineering Research
• /
• v.44 no.5
• /
• pp.346-353
• /
• 2023

In this paper, we propose an innovative approach that leverages deep learning to find optimal reference points for achieving precise tooth segmentation in three-dimensional tooth point cloud data. A dataset consisting of 350 aligned maxillary and mandibular cloud data was used as input, and both end coordinates of individual teeth were used as correct answers. A two-dimensional image was created by projecting the rendered point cloud data along the Z-axis, where an image of individual teeth was created using an object detection algorithm. The proposed algorithm is designed by adding various modules to the Unet model that allow effective learning of a narrow range, and detects both end points of the tooth using the generated tooth image. In the evaluation using DSC, Euclid distance, and MAE as indicators, we achieved superior performance compared to other Unet-based models. In future research, we will develop an algorithm to find the reference point of the point cloud by back-projecting the reference point detected in the image in three dimensions, and based on this, we will develop an algorithm to divide the teeth individually in the point cloud through image processing techniques.

• Korean Journal of Ecology and Environment
• /
• v.43 no.2
• /
• pp.319-337
• /
• 2010

Major objective of this study was to determine interannual and seasonal water quality along with characteristics of longitudinal gradients along the reservoir axis of the riverine zone (Rz)-to-lacustrine zone (Lz). Water quality dataset of five years during 2003~2007 used here were obtained from Ministry of Environment, Korea and ten physical, chemical and biological parameters were analyzed in the study. Similarity analysis, based on moropho-hydrological variables of reservoir surface area, watershed area, total inflow, and outflow, showed that the reservoirs were categorized as three groups of large-dam reservoirs (Chungju Reservoir, Daecheong Reservoir and Soyang Reservoir), mid-size reservoirs (Andong Reservoir, Yongdam Reservoir, Juam Reservoir and Hapcheon Reservoir), and small-size reservoirs (Hoengseong Reservoir and Buan Reservoir). According to the data comparison of high-flow year (2003) vs. lowflow year (2005), dissolved oxygen (DO), pH, biological oxygen demand (BOD), suspended solids (SS), total nitrogen (TN), total phosphorus (TP), chlorophyll-a (CHL) and electrical conductivity (EC) declined along the longitudinal axis of Rz to Lz and water transparency, based on Secchi depth (SD), increased along the axis. These results indicate that transparency was a function of Values of pH, DO, SS, SD, and EC at each site were greater in the low-flow year (2005) than the high-flow year (2003), whereas values of BOD, COD, TN, TP and CHL were greater in the high-flow year (2003). When values of TN, TP, CHL and SD in nine reservoirs were compared in the three zones of Rz, Tz, and Lz, values of TN, TP and CHL declined along longitudinal gradients and SD showed the opposite due to the sedimentation processes from the water column. Values of TN were not statistically correlated with TP values. The empirical linear models of TP-CHL and CHL-SD showed significant (p<0.05, $R^2$>0.04). In the mid-size reservoirs, the variation of CHL was explained ($R^2$=0.2401, p<0.0001, n=239) by the variation of TP. The affinities in the correlation analysis of mid-size reservoirs were greater in the CHL-SD model than any other empirical models, and the CHL-SD model had an inverse relations. In the meantime, water quality variations was evidently greater in Daecheong Reservoir than two reservoirs of Andong Reservoir and Hoengseong Reservoir as a result of large differences of water quality by long distance among Rz, Tz and Lz.

• Korean Journal of Ecology and Environment
• /
• v.40 no.2
• /
• pp.184-192
• /
• 2007

This study was to evaluate ecological health of Sumjin River during April${\sim}$June 2006. The ecological health assessments was based on the Index of Biological Integrity (IBI), Qualitative Babitat Evaluation Index (QHEI), and water chemistry. For the study, the models of IBI and QHEI were modified as 10 and 11 metric attributes, respectively. We also analyzed spatial patterns of chemical water quality over the period of $2002{\sim}2005$, using the water chemistry dataset, obtained from the Ministry of Environment, Korea. In Sumjin River, values of IBI averaged 33 (n= 12), which is judged as a "Fair${\sim}$Good" condition after the criteria of Barbour at al. (1999). There was a distinct spatial variation. Mean IBI score at Site 5 was estimated as 40, indicating a "Good" condition whereas, the mean at Site 3 was 23, indicating a "Poor${\sim}$Fair" condition. Habitat analysis showed that QHEI values in the river averaged 109 (n=6), indicating a "Marginal" condition after the criteria of Harbour et al. (1999). Values of BOD and COD averaged 1.3 mg $L^{-1}$ (scope: $0.9{\sim}1.8$ mg $L^{-1}$) and 3.3 mg $L^{-1}$ (scope: $2.8{\sim}4.0$ mg $L^{-1}$), respectively during the study. It was evident that chemical pollutions by organic matter were minor in the river. Total nitrogen (TN) and total phosphorus (TP) averaged 2.5 mg $L^{-1}$ and 0.067 mg $L^{-1}$, respectively, and the nutrients did not show large longitudinal gradients between the upper and lower reach. Overall, dataset of IBI, QHEI, and water chemistry suggest that river health has been well maintained, compared to other major watersheds in Korea and should be protected from habitat disturbance and chemical pollutions.

• Korean Journal of Ecology and Environment
• /
• v.43 no.3
• /
• pp.428-436
• /
• 2010

This study was to evaluate water quality characteristics and ecological health using a mulimetric fish model in Taehwa-River watershed during May~September 2009. The ecological health assessments were based on the Index of Biological Integrity (IBI) using fish community and the multimetric model of Qualitative Habitat Evaluation Index (QHEI). For the study, the models of IBI and QHEI were modified as 8 and 11 metric attributes, respectively. We also analyzed spatial patterns of chemical water quality over the period of 2000~2009, using the water chemistry dataset, obtained from the Ministry of Environment, Korea. Values of BOD and COD averaged $1.7\;mg\;L^{-1}$ (scope: $0.1{\sim}31.8\;mg\;L^{-1}$) and $3.6\;mg\;L^{-1}$ (scope: $0.4{\sim}33\;mg\;L^{-1}$), respectively during the study. Total nitrogen (TN) and total phosphorus (TP) averaged $2.8\;mg\;L^{-1}$ and $96.8\;{\mu}g\;L^{-1}$, respectively, indicating an eutrophic-hypertrophic state. Also, TN and TP showed longitudinal increases toward the downriver reach. In the watershed, QHEI values varied from 67.5 (fair condition) to 164.5 (good condition) by the criteria of US EPA (1993). There was a abruptly decreasing tendency from T9 site in the QHEI values. According to 1st and 2nd surveys of Taewha River, multimetric model values of IBI was averaged 26.1 (n=14) with "good" condition (B) and the spatial variation was evident. Our results suggest that the mainstream sites was getting worse health condition along the river gradient due to inputs of the point and non-point sources from the urban (Ulsan city). Overall, dataset of IBI, QHEI, and water chemistry indicated that the ecological river health showed a downriver decline and the pattern was closely associated with habitat degradations and chemical pollutions as the waters pass through the urban region.

• Korean Journal of Ecology and Environment
• /
• v.40 no.1
• /
• pp.72-81
• /
• 2007

The purpose of this study was to develop a multi-metric Lentic Ecosystem Health Assessment (LEHA) model and apply model to dataset sampled from Daechung Reservoir in September 2005. The metrics were composed of 11 parameters such as physical, chemical and biological variables. The metric attributes of $M_1{\sim}M_8$ followed after the model of biological integrity using fish assemblages that previously adapted in lotic ecosystems, while the metrics of $M_9{\sim}M_{11}$ were added on the basis of literature. The metric of $M_9$ reflected habitat conditions in the littoral zone and the metric of $M_{10}$ reflected chemical conditions of the reservoir. For the application of regression analysis of long-transformed conductivity [$Log_{10}$(Cond)] against $COD_{Mn}$, based on 150 sampling sites at Korean reservoirs, showed that the variation of conductivity was explained 77.4% [$COD_{Mn}=4.42{\times}Log_{10}(Cond)-5.43;\;R^2=0.774$, p<0.01, n=150] by the variation of $COD_{Mn}$. The metric of $M_{11}$ was based on Tropic State Index (TSI), based on chlorophyll-${\alpha}$ concentrations (Chl-${\alpha}$). Analysis of TSI $(Chl-{\alpha})$ showed that above 50 was estimated "1", $40{\sim}50$ was estimated "3" and below 40% was estimated '5'. Overall, velues of LEHA in the reservoir averaged 30.5, indicating a "fair${\sim}$poor condition", which is judged by the criteria of U.S. EPA (1993). More studies such as metric numbers and attributes should be done for the application of the model.

• Korean Journal of Ecology and Environment
• /
• v.41 no.2
• /
• pp.228-236
• /
• 2008

The purpose of this study was to evaluate fish metric attributes using a model of Lentic Ecosystem Health Assessment (LEHA) and apply the model to the dataset sampled from six sites of Namyang Reservoir during October 2005$\sim$May 2006. The model was composed of 11 metries and the metric attributes were made of physical, chemical and biological parameters. Trophic composition's metrics showed that tolerant species ($M_3$, 80%) and omnivore species ($M_4$, 92%) dominated the fish fauna, indicating a biological degradation in the aquatic ecosystem. The metric of $M_7$, relative proportions of exotic species, also showed greater than 8% of the total, indicating a ecological disturbance. The average value of LEHA model was 24.3 (n= 12) in the reservoir, indicating a "poor condition" by the criteria of An and Han (2007). Spatial variation based on the model values was low (range: $21{\sim}26$), and temporal variation occurred due to a monsoon rainfall. Electrical conductivity (EC) and tropic state index of chlorophyll-$\alpha$ [TSI(CHL)] was greater in the premonsoon than the postmonsoon.

• Horticultural Science & Technology
• /
• v.29 no.2
• /
• pp.116-122
• /
• 2011

In this study, we report the generation and analysis of 1,392 expressed sequence tags (ESTs) from Korean Stewartia (Stewartia koreana Nakai). A cDNA library was generated from the young leaf tissue and a total of 1,392 cDNA were partially sequenced. EST and unigene sequence quality were determined by computational filtering, manual review, and BLAST analyses. Finally, 1,301 ESTs were acquired after the removal of the vector sequence and filtering over a minimum length 100 nucleotides. A total of 893 unigene, consisting of 150 contigs and 743 singletons, was identified after assembling. Also, we identified 95 new microsatellite-containing sequences from the unigenes and classified the structure according to their repeat unit. According to homology search with BLASTX against the NCBI database, 65% of ESTs were homologous with known function and 11.6% of ESTs were matched with putative or unknown function. The remaining 23.2% of ESTs showed no significant similarity to any protein sequences found in the public database. Annotation based searches against multiple databases including wine grape and populus sequences helped to identify putative functions of ESTs and unigenes. Gene ontology (GO) classification showed that the most abundant GO terms were transport, nucleotide binding, plastid, in terms biological process, molecular function and cellular component, respectively. The sequence data will be used to characterize potential roles of new genes in Stewartia and provided for the useful tools as a genetic resource.

• The KIPS Transactions:PartB
• /
• v.13B no.7 s.110
• /
• pp.679-688
• /
• 2006

Protein-protein interaction data obtained from high-throughput experiments includes high false positives. In this paper, we introduce a new protein-protein interaction reliability verification system. The proposed system integrates various biological features related with protein-protein interactions, and then selects the most relevant and informative features among them using a feature selection method. To assess the reliability of each protein-protein interaction data, the system construct a classifier that can distinguish true interacting protein pairs from noisy protein-protein interaction data based on the selected biological evidences using a classification technique. Since the performance of feature selection methods and classification techniques depends heavily upon characteristics of data, we performed rigorous comparative analysis of various feature selection methods and classification techniques to obtain optimal performance of our system. Experimental results show that the combination of feature selection method and classification algorithms provide very powerful tools in distinguishing true interacting protein pairs from noisy protein-protein interaction dataset. Also, we investigated the effects on performances of feature selection methods and classification techniques in the proposed protein interaction verification system.