• 제목/요약/키워드: Focal loss

검색결과 122건 처리시간 0.026초

Effects of Melatonin on Improvement of Neurological Function in Focal Cerebral Ischemic Rats

  • Lee, Seung-Hoon;Shin, Jin-Hee;Lee, Min-Kyung;Lee, Sang-Kil;Lee, Sang-Rae;Chang, Kyu-Tae;Hong, Yong-Geun
    • Reproductive and Developmental Biology
    • /
    • 제35권2호
    • /
    • pp.167-174
    • /
    • 2011
  • Acute ischemic stroke results from sudden decrease or loss of blood supply to an area of the brain, resulting in a coinciding loss of neurological function. The antioxidant action of melatonin is an important mechanism among its known effects to protective activity during ischemic/reperfusion injury. The focus of this research, therapeutic efficacy of melatonin on recovery of neurological function following long term treatment in ischemic brain injured rats. Male Sprague-Dawley rats (n=40; 8 weeks old) were divided into the control group, and MCAo groups (Vehicle, MT7 : MCAo+ melatonin injection at 7:00, MT19 : MCAo+melatonin injection at 19:00, and MT7,19 : MCAo+melatonin injection at 7:00 and 19:00). Rat body weight and neurological function were measured every week for 8 weeks. After 8 weeks, the rats were anesthetized with a mixture of zoletil (40 mg/kg) and xylazine (10 mg/kg) and sacrificed for further analysis. Tissues were then collected for RNA isolation from brain tissue. Also, brain tissues were analyzed by histological procedures. We elucidated that melatonin was not toxic in vital organs. MT7,19 was the most rapidly got back to mild symptom on test of neurological parameter. Also, exogenous melatonin induces both the down-regulation of detrimental genes, such as NOSs and the up-regulation of beneficial gene, including BDNF during long term administration after focal cerebral ischemia. Melatonin treatment reduced the loss of primary motor cortex. Therefore, we suggest that melatonin could be act as prophylactic as well as therapeutic agent for neurorehabilitative intervention.

흉부 CT 영상에서 심층 감독 및 하이브리드 병변 초점 손실 함수를 활용한 폐암 분할 개선 (Enhanced Lung Cancer Segmentation with Deep Supervision and Hybrid Lesion Focal Loss in Chest CT Images)

  • 이민진;오윤선;홍헬렌
    • 한국컴퓨터그래픽스학회논문지
    • /
    • 제30권1호
    • /
    • pp.11-17
    • /
    • 2024
  • 폐암은 크기가 다양하고 유사한 밝기값을 갖는 주변 구조물이 존재하기 때문에 흉부 CT 영상에서 폐암을 정확하게 분할하는 것이 어렵다. 이러한 문제를 해결하기 위해 본 논문에서는 심층 감독을 포함하고 UNet3+를 백본으로 사용하는 폐암 분할 네트워크를 제안한다. 또한, 픽셀 기반, 영역 기반 및 형태 기반의 3가지 구성 요소로 이루어진 하이브리드 병변 초점 손실함수를 제안한다. 이를 통해 배경에 비해 작은 영역을 차지하는 폐암 부분에 집중하고, 불명확한 경계를 처리하는데 도움이 되는 형태 정보를 고려할 수 있다. 제안 방법을 UNet 및 UNet3+와 비교 실험을 통해 검증하였고, 제안 방법은 모든 폐암 크기에서 DSC 측면에서 가장 우수한 성능을 보였다.

Compound Loss Function of semantic segmentation models for imbalanced construction data

  • Chern, Wei-Chih;Kim, Hongjo;Asari, Vijayan;Nguyen, Tam
    • 국제학술발표논문집
    • /
    • The 9th International Conference on Construction Engineering and Project Management
    • /
    • pp.808-813
    • /
    • 2022
  • This study presents the problems of data imbalance, varying difficulties across target objects, and small objects in construction object segmentation for far-field monitoring and utilize compound loss functions to address it. Construction site scenes of assembling scaffolds were analyzed to test the effectiveness of compound loss functions for five construction object classes---workers, hardhats, harnesses, straps, hooks. The challenging problem was mitigated by employing a focal and Jaccard loss terms in the original loss function of LinkNet segmentation model. The findings indicates the importance of the loss function design for model performance on construction site scenes for far-field monitoring.

  • PDF

재관류시 외인성 칼슘 자극이 허혈 심근에 미치는 영향에 대한 전자현미경적 연구 (An Ultrastructural Study on the Effect of Exogenous $Ca^{2+}$ Stimulation to Ischemic Myocardium during Post-ischemic Reperfusion)

  • 김호덕;전상배;라봉진
    • Applied Microscopy
    • /
    • 제21권1호
    • /
    • pp.1-20
    • /
    • 1991
  • The effects of exogenous $Ca^{2+}$ stimulation on the post-ischemic myocardial cells were studied using isolated Langendorff-perfused guinea pig hearts. At the starting point of reperfusion, Tyrode solutions, each containing 2.0mM, 4.0mM and 8.0mM $CaCl_2$ respectively, were administered for 2 minutes apart by descending, ascending, or by combined sequences followed by standard Tyrode solution containing 1.0mM $CaCl_2$. The minutes of global ischemia produced reversible but moderate to severe degree of myocardial ultrastructrual changes including focal destruction of sarcolemma, loss of nuclear matrix, clumping and margination of chromatins, mitochondrial swelling, destruction of mitochondrial cristae, shortening of sarcomeres, focal loss of myofibrils, and separation of cell junctions. In spite of reperfusion, the ultrastructure was more severely damaged and irreversible changes such as intracellular fluid accumulation, contracted sarcomeres, mitochondrial destruction, disruption of sarcolemma, loss of nuclear matrix, and separation of cell junction were observed in a large number of cells. In contrast, Tyrode-perfused $Ca^{2+}$-stimulated myocardial cells showed relatively well preserved ultrastucture, except slight changes including focal mitochondrial swelling, widening of T-tubule, and widening of cell junctions, especially at fasciae adherentes. The post-ischemic $Ca^{2+}$-stimulated reperfused myocardial cells produced focal changes such as mitochondrial destruction, disintegration of sarcolemma, widening of T-tubule, and intracellular fluid accumulation with slight variation in degree of changes by the method of $Ca^{2+}$ administration sequence. However, in a large number of the myocardial cells, chromatins were redistributed relatively evenly in the nuclear matrix, mitochondrial cristae were tightly packed, and a considerable number of intramitochondrial granules and glycogen granules reap-pealed. These results indicate that exogenous $Ca^{2+}$ stimulation in the initial period of reperfusion may be beneficial to salvage or to reduce the post-ischemic myocardium from further deleterious changes, and that the beneficial effects may be derived from the reserves of the function of the intracellular $Ca^{2+}$ regulating organelles and/or from the responsiveness of contractile apparatus to $Ca^{2+}$ stimulation.

  • PDF

진행성 치주염이 지수 조직에 미치는 영향 (The influence of Advanced Adult Periodontitis on the pulp)

  • 이강운;이철우;한수부
    • Journal of Periodontal and Implant Science
    • /
    • 제29권1호
    • /
    • pp.95-102
    • /
    • 1999
  • The purpose of this study was to observe histopathologically the influence of advanced periodontitis on pulp tissue, and to conclude the correlation between the results with clinical manifestations. The samples were teeth with over 7mm pocket depth and over 50% radiographic bone loss. These were diagnosed to have very poor prognosis and thus planned to be extracted. Those with any of following conditions were excluded from the samples, loss of vitality, periapical pathology, restoration or prosthesis, dental caries, and attrition or abrasion. It was because these conditions could affect pulp without any correlation with periodontal disease. For the experiment, 17 teeth from 11 patients were selected. Average age of patient was 47. Each tooth was examined for following categoris; pocket depth, gingival recession, electric pulp test, mobility, percussion test, sensitivity test. The extracted teeth were fixed buffered neutral formalin solution. It was decalcified using 4% nitric acid. Sliced histological samples observed using light microscope, for pulp status, and severeity of inflammation. 4 samples were excluded due to histologic sample discrepency. Thus 13 samples were subject to observation. 4 showed normal conditions. Focal reversable pulpitis was shown in 5 samples. Chronic pulpitis was observed 1 sample. Pulpal abscess was observed in 3 samples.

  • PDF

Wide band prototype feedhorn design for ASTE focal plane array

  • Lee, Bangwon;Gonzales, Alvaro;Lee, Jung-won
    • 천문학회보
    • /
    • 제41권2호
    • /
    • pp.66.2-66.2
    • /
    • 2016
  • KASI and NAOJ are making collaborating efforts to implement faster mapping capability into the new 275-500 GHz Atacama Submillimeter Telescope Experiment focal plane array (FPA). Feed horn antenna is one of critical parts of the FPA. Required fractional bandwidth is almost 60 % while that of traditional conical horn is less than 50 %. Therefore, to achieve this wideband performance, we adopted a horn of which the corrugation depths have a longitudinal profile. A profiled horn has features not only of wide bandwidth but also of shorter length compared to a linear-tapered corrugated horn, and lower cost fabrication with less error can be feasible. In our design process the flare region is represented by a cubic splined curve with several parameters. Parameters of the flare region and each dimension of the throat region are optimized by a differential evolution algorithm to keep >20 dB return loss and >30 dB maximum cross-polarization level over the operation bandwidth. To evaluate RF performance of the horn generated by the optimizer, we used a commercial mode matching software, WASP-NET. Also, Gaussian beam (GB) masks to far fields were applied to give better GB behavior over frequencies. The optimized design shows >23 dB return loss and >33 dB maximum cross-polarization level over the whole band. Gaussicity of the horn is over 96.6 %. The length of the horn is 12.5 mm which is just 57 % of the ALMA band 8 feed horn (21.96 mm).

  • PDF

Convolutional neural network-based data anomaly detection considering class imbalance with limited data

  • Du, Yao;Li, Ling-fang;Hou, Rong-rong;Wang, Xiao-you;Tian, Wei;Xia, Yong
    • Smart Structures and Systems
    • /
    • 제29권1호
    • /
    • pp.63-75
    • /
    • 2022
  • The raw data collected by structural health monitoring (SHM) systems may suffer multiple patterns of anomalies, which pose a significant barrier for an automatic and accurate structural condition assessment. Therefore, the detection and classification of these anomalies is an essential pre-processing step for SHM systems. However, the heterogeneous data patterns, scarce anomalous samples and severe class imbalance make data anomaly detection difficult. In this regard, this study proposes a convolutional neural network-based data anomaly detection method. The time and frequency domains data are transferred as images and used as the input of the neural network for training. ResNet18 is adopted as the feature extractor to avoid training with massive labelled data. In addition, the focal loss function is adopted to soften the class imbalance-induced classification bias. The effectiveness of the proposed method is validated using acceleration data collected in a long-span cable-stayed bridge. The proposed approach detects and classifies data anomalies with high accuracy.

A Case of Secondary FSGS due to Chronic Chloride Diarrhea

  • Kim, Byung Kwan;Lee, Hyun Soon;Yim, Hyung Eun;Cheong, Hae Il;Yoo, Kee Hwan
    • Childhood Kidney Diseases
    • /
    • 제20권2호
    • /
    • pp.83-87
    • /
    • 2016
  • Congenital chloride diarrhea (CLD) is a rare autosomal recessive disease that is difficult to diagnose. CLD requires early treatment to correct electrolyte imbalance and alkalosis and to prevent severe dehydration. Renal injury is clearly associated with defective electrolyte balance induced by CLD, particularly during the first months or years of life. A 7-year-old boy was diagnosed with CLD following detection of a homozygous mutation (c.2063-1G>T) in SLC26A3 at 6 months of age. During treatment with electrolyte supplements, mild proteinuria was detected at 8 months of age, and is still present. Renal biopsy showed the presence of focal renal dysplasia, with metaplastic cartilage and mononuclear cell infiltration, calcification, and fibrosis in the interstitium. Up to two-thirds of the glomeruli exhibited global obsolescence, mostly aggregated in the dysplastic area. In nondysplastic areas, the glomeruli were markedly increased in size and severely hypercellular, with increased mesangial matrix, and displayed segmental sclerosis. The marked glomerular hypertrophy with focal segmental glomerulosclerosis suggested a compensatory reaction to the severe nephron loss or glomerular obsolescence associated with renal dysplasia, with superimposed by CLD aggravating the tubulointerstitial damage.

흉부 볼륨 CT영상에서 Weighted Integration Loss을 이용한 폐암 분할 알고리즘 연구 (A Study on Lung Cancer Segmentation Algorithm using Weighted Integration Loss on Volumetric Chest CT Image)

  • 정진교;김영재;김광기
    • 한국멀티미디어학회논문지
    • /
    • 제23권5호
    • /
    • pp.625-632
    • /
    • 2020
  • In the diagnosis of lung cancer, the tumor size is measured by the longest diameter of the tumor in the entire slice of the CT. In order to accurately estimate the size of the tumor, it is better to measure the volume, but there are some limitations in calculating the volume in the clinic. In this study, we propose an algorithm to segment lung cancer by applying a custom loss function that combines focal loss and dice loss to a U-Net model that shows high performance in segmentation problems in chest CT images. The combination of values of the various parameters in custom loss function was compared to the results of the model learned. The purposed loss function showed F1 score of 88.77%, precision of 87.31%, recall of 90.30% and average precision of 0.827 at α=0.25, γ=4, β=0.7. The performance of the proposed custom loss function showed good performance in lung cancer segmentation.

One-step deep learning-based method for pixel-level detection of fine cracks in steel girder images

  • Li, Zhihang;Huang, Mengqi;Ji, Pengxuan;Zhu, Huamei;Zhang, Qianbing
    • Smart Structures and Systems
    • /
    • 제29권1호
    • /
    • pp.153-166
    • /
    • 2022
  • Identifying fine cracks in steel bridge facilities is a challenging task of structural health monitoring (SHM). This study proposed an end-to-end crack image segmentation framework based on a one-step Convolutional Neural Network (CNN) for pixel-level object recognition with high accuracy. To particularly address the challenges arising from small object detection in complex background, efforts were made in loss function selection aiming at sample imbalance and module modification in order to improve the generalization ability on complicated images. Specifically, loss functions were compared among alternatives including the Binary Cross Entropy (BCE), Focal, Tversky and Dice loss, with the last three specialized for biased sample distribution. Structural modifications with dilated convolution, Spatial Pyramid Pooling (SPP) and Feature Pyramid Network (FPN) were also performed to form a new backbone termed CrackDet. Models of various loss functions and feature extraction modules were trained on crack images and tested on full-scale images collected on steel box girders. The CNN model incorporated the classic U-Net as its backbone, and Dice loss as its loss function achieved the highest mean Intersection-over-Union (mIoU) of 0.7571 on full-scale pictures. In contrast, the best performance on cropped crack images was achieved by integrating CrackDet with Dice loss at a mIoU of 0.7670.