Your browser doesn't support javascript.
loading
Show: 20 | 50 | 100
Results 1 - 7 de 7
Filter
Add more filters










Database
Language
Publication year range
1.
Comput Biol Med ; 157: 106750, 2023 05.
Article in English | MEDLINE | ID: mdl-36931202

ABSTRACT

Diabetic retinopathy(DR) is a common early diabetic complication and one of the main causes of blindness. In clinical diagnosis and treatment, regular screening with fundus imaging is an effective way to prevent the development of DR. However, the regular fundus images used in most DR screening work have a small imaging range, narrow field of vision, and can not contain more complete lesion information, which leads to less ideal automatic DR grading results. In order to improve the accuracy of DR grading, we establish a dataset containing 101 ultra-wide-field(UWF) DR fundus images and propose a deep learning(DL) automatic classification method based on a new preprocessing method. The emerging UWF fundus images have the advantages of a large imaging range and wide field of vision and contain more information about the lesions. In data preprocessing, we design a data denoising method for UWF images and use data enhancement methods to improve their contrast and brightness to improve the classification effect. In order to verify the efficiency of our dataset and the effectiveness of our preprocessing method, we design a series of experiments including a variety of DL classification models. The experimental results show that we can achieve high classification accuracy by using only the backbone model. The most basic ResNet50 model reaches an average of classification accuracy(ACA) 0.66, Macro F1 0.6559, and Kappa 0.58. The best-performing Swin-S model reaches ACA 0.72, Macro F1 0.7018, and Kappa 0.65. DR grading using UWF images can achieve higher accuracy and efficiency, which has practical significance and value in clinical applications.


Subject(s)
Diabetes Mellitus , Diabetic Retinopathy , Humans , Diabetic Retinopathy/diagnostic imaging , Fundus Oculi , Photography/methods
2.
Org Lett ; 24(44): 8192-8196, 2022 Nov 11.
Article in English | MEDLINE | ID: mdl-36318750

ABSTRACT

A bromoalkane-directed radical 1,4-aryl shift strategy for nickel-catalyzed reductive Heck-type C(sp3)-C(sp2) coupling cascades of α-amino-ß-bromocarboxylic acid esters with α-trifluoromethyl alkenes for producing gem-difluorinated arylalanines is presented. The α-aminoalkyl radicals generated from neophyl-type aryl migration function as robust coupling partners to allow for further Giese-type addition with electron-deficient α-trifluoromethyl alkenes and vinyl sulfones, thereby realizing a new radical cascade for the simultaneous installation of an aromatic ring and olefin motif into amino acid backbones.

3.
IEEE Trans Pattern Anal Mach Intell ; 43(5): 1791-1807, 2021 05.
Article in English | MEDLINE | ID: mdl-31751226

ABSTRACT

State-of-the-art classification models are usually considered as black boxes since their decision processes are implicit to humans. On the contrary, human experts classify objects according to a set of explicit hierarchical criteria. For example, "tabby is a domestic cat with stripes, dots, or lines", where tabby is defined by combining its superordinate category (domestic cat) and some certain attributes (e.g., has stripes). Inspired by this mechanism, we propose an interpretable Hierarchical Criteria Network (HCN) by additionally learning such criteria. To achieve this goal, images and semantic entities (e.g., taxonomies and attributes) are embedded into a common space, where each category can be represented by the linear combination of its superordinate category and a set of learned discriminative attributes. Specifically, a two-stream convolutional neural network (CNN) is elaborately devised, which embeds images and taxonomies with the two streams respectively. The model is trained by minimizing the prediction error of hierarchy labels on both streams. Extensive experiments on two widely studied datasets (CIFAR-100 and ILSVRC) demonstrate that HCN can learn meaningful attributes as well as reasonable and interpretable classification criteria. Therefore, the proposed method enables further human feedback for model correction as an additional benefit.

4.
Org Lett ; 21(17): 6761-6764, 2019 09 06.
Article in English | MEDLINE | ID: mdl-31433660

ABSTRACT

Total synthesis of (±)-cochlearol A was accomplished, which features a cis 6/6 B/D ring synthesis. A TMSOTf-promoted lactonization of tert-butoxy ketoester produced the desired lactone with quaternary carbon. The cis configuration of the B/E ring is essential for regioselective B/D ring formation. Finally, simple deprotections and transformations gave cochlearol A in 16 steps from known ethyl 4-tert-butoxyacetoacetate.

5.
Chem Commun (Camb) ; 52(55): 8561-4, 2016 Jun 30.
Article in English | MEDLINE | ID: mdl-27321202

ABSTRACT

Enantioselective total synthesis of (+)-Lingzhiol has been achieved. It is the first example of in tandem semipinacol rearrangement reactions, the migrated aryl group further reacting with the carbonyl oxonium electrophile to furnish a polycyclic skeleton. Our synthesis involves 13 steps and proceeds in 6% overall yield.

6.
Chem Commun (Camb) ; 51(78): 14594-6, 2015 Oct 07.
Article in English | MEDLINE | ID: mdl-26288857

ABSTRACT

Concise synthesis of (±)-Lingzhiol has been achieved. The key reaction involves one-step construction of a 5/5/6/6 tetra-ring backbone of Lingzhiol via epoxy-arene cyclization.


Subject(s)
Epoxy Compounds/chemistry , Terpenes/chemical synthesis , Cyclization , Stereoisomerism , Terpenes/chemistry
7.
Chem Biol Drug Des ; 84(3): 364-77, 2014 Sep.
Article in English | MEDLINE | ID: mdl-24674599

ABSTRACT

Inhibition of dipeptidyl peptidase IV is an important approach for the treatment of type-2 diabetes. In this study, we reported a multistage virtual screening workflow that integrated 3D pharmacophore models, structural consensus docking, and molecular mechanics/generalized Born surface area binding energy calculation to identify novel dipeptidyl peptidase IV inhibitors. After screening our in-house database, two hit compounds, HWL-405 and HWL-892, having persistent high performance in all stages of virtual screening were identified. These two hit compounds together with several analogs were synthesized and evaluated for in vitro inhibition of dipeptidyl peptidase IV. The experimental data indicated that most designed compounds exhibited significant dipeptidyl peptidase IV inhibitory activity. Among them, compounds 35f displayed the greatest potency against dipeptidyl peptidase IV in vitro with the IC50 value of 78 nm. In an oral glucose tolerance test in normal male Kunming mice, compound 35f reduced blood glucose excursion in a dose-dependent manner.


Subject(s)
Dipeptidyl-Peptidase IV Inhibitors/chemical synthesis , Hypoglycemic Agents/chemical synthesis , Animals , Area Under Curve , Binding Sites , Databases, Chemical , Diabetes Mellitus, Type 2/drug therapy , Dipeptidyl Peptidase 4/chemistry , Dipeptidyl Peptidase 4/metabolism , Dipeptidyl-Peptidase IV Inhibitors/metabolism , Dipeptidyl-Peptidase IV Inhibitors/therapeutic use , Drug Evaluation, Preclinical , Glucose Tolerance Test , Hypoglycemic Agents/chemistry , Hypoglycemic Agents/therapeutic use , Male , Mice , Molecular Docking Simulation , Protein Structure, Tertiary , ROC Curve , Structure-Activity Relationship
SELECTION OF CITATIONS
SEARCH DETAIL
...