Your browser doesn't support javascript.
loading
Show: 20 | 50 | 100
Results 1 - 3 de 3
Filter
Add filters








Language
Year range
1.
Journal of Southern Medical University ; (12): 815-824, 2023.
Article in Chinese | WPRIM | ID: wpr-986993

ABSTRACT

OBJECTIVE@#We propose a novel region- level self-supervised contrastive learning method USRegCon (ultrastructural region contrast) based on the semantic similarity of ultrastructures to improve the performance of the model for glomerular ultrastructure segmentation on electron microscope images.@*METHODS@#USRegCon used a large amount of unlabeled data for pre- training of the model in 3 steps: (1) The model encoded and decoded the ultrastructural information in the image and adaptively divided the image into multiple regions based on the semantic similarity of the ultrastructures; (2) Based on the divided regions, the first-order grayscale region representations and deep semantic region representations of each region were extracted by region pooling operation; (3) For the first-order grayscale region representations, a grayscale loss function was proposed to minimize the grayscale difference within regions and maximize the difference between regions. For deep semantic region representations, a semantic loss function was introduced to maximize the similarity of positive region pairs and the difference of negative region pairs in the representation space. These two loss functions were jointly used for pre-training of the model.@*RESULTS@#In the segmentation task for 3 ultrastructures of the glomerular filtration barrier based on the private dataset GlomEM, USRegCon achieved promising segmentation results for basement membrane, endothelial cells, and podocytes, with Dice coefficients of (85.69 ± 0.13)%, (74.59 ± 0.13)%, and (78.57 ± 0.16)%, respectively, demonstrating a good performance of the model superior to many existing image-level, pixel-level, and region-level self-supervised contrastive learning methods and close to the fully- supervised pre-training method based on the large- scale labeled dataset ImageNet.@*CONCLUSION@#USRegCon facilitates the model to learn beneficial region representations from large amounts of unlabeled data to overcome the scarcity of labeled data and improves the deep model performance for glomerular ultrastructure recognition and boundary segmentation.


Subject(s)
Humans , Electrons , Endothelial Cells , Learning , Podocytes , Kidney Diseases
2.
Acta Pharmaceutica Sinica ; (12): 1057-63, 2010.
Article in Chinese | WPRIM | ID: wpr-382322

ABSTRACT

The aim of this study is to investigate the critical factor affecting the properties of PLGA microspheres fabricated by a solid-in-oil-in-water (S/O/W) emulsion technique with BSA as a model protein. Prior to encapsulation, the BSA microparticles were fabricated by a modified freezing-induced phase separation method. The microparticles were subsequently encapsulated into PLGA microspheres by S/O/W emulsion method, then Motic BA200 biological microscope, confocal laser scanning microscope, scanning electron microscope were used to observe the structure of S/O/W emulsion and PLGA microspheres. The protein content extracted or released from BSA microspheres was measured by Bradford protein assay method. It was found that NaCl added in the outer aqueous phase effectively suppressed material exchange between the inner and outer phase of S/O/W emulsion. Then, the structure and permeability of obtained microspheres were influenced. As a result, with the increase of NaCl concentration in the outer aqueous phase, the encapsulation efficiency of microspheres significantly increased from 60% to more than 85%, the burst release of microspheres reduced from 70% to 20%, and the particle size decreased from 103 microm to 62 microm. Furthermore, the rehydration of encapsulated protein was also retarded and then integrity of BSA was successfully protected during encapsulation process. In vitro release test showed that BSA released from PLGA microspheres in a sustained manner for more than 30 days.

3.
Chinese Journal of Tissue Engineering Research ; (53): 8931-8934, 2008.
Article in Chinese | WPRIM | ID: wpr-406845

ABSTRACT

BACKGROUND: The core of bone tissue engineering is to construct a scaffold that is similar to human bone tissue structure and features.OBJECTIVE: To compare pathochemical and mechanical characteristics between pig and human bone scaffold materials.DESIGN, TIME AND SETTING: Contrast study was performed at Clinical Anatomy Institute, South Medical University; Guangdong Province Key Laboratory of Tissue Construction and Detection from March to December 2006.MATERIALS: Four fresh health adult human cadavers were provided by South Medical University, Guangzhou Red Cross Society, and the relatives knew the fact. Ultra low temperature freezing 6-month iliac bones of 6 adult swines were also used in this study.METHODS: Pig iliac and healthy adults iliac bones were obtained to remove soft tissue, curettage periosteum and bone marrow. Bone sawing machine was used to cut cancellous bone into smaller bone sections around 5 mm×5 mm×40 mm, which underwent ultrasonic cleaning, H2O2 and alcohol soaking, freeze drying and radiation treatment; finally, xenogeneic bone scaffold and allogeneic bone scaffold were obtained.MAIN OUTCOME MEASURES: Xenogeneic bone scaffold material and human allograft bone scaffold were observed with scanning electron microscopy to compare porosity, contents of protein content, calcium and phosphorus, and mechanical properties.RESULTS: Xenogeneic bone scaffold and allogeneic bone scaffold both had intrinsical bone trabecula, trabecular spaces and bone cavity system. Both of them had unabridged natural three dimensional network structure. The 3D supporting frames of them were complete. The xenogeneic bone scaffold had more spaces than allogeneic bone scaffold. The size of both scaffolds was approximation, about 400 μm. The interval porosity of xenogeneic bone scaffold was higher than the allogeneic bone scaffold (P<0.05). And the protein of xenogeneic bone scaffold was not as many as it of allogeneic bone scaffold (P<0.05). The contents of Ca and P were similar (P>0.05), and there was no significant difference in Young's modulus of xenogeneic bone scaffold and allogeneic bone scaffold (P>0.05).CONCLUSION: Xenogeneic bone scaffold may completely meet the clinical demands for bone grafting or be the scaffold of bone tissue engineering in mechanical chemical properties.

SELECTION OF CITATIONS
SEARCH DETAIL