Conventional ultrasound (US) has been routinely used for differential diagnosis of thyroid nodules, but its discriminatory performance remains unsatisfactory. This study aimed to develop and validate a prediction nomogram model based on conventional US contrast-enhanced (CEUS) features differentiating malignant from benign nodules. A total 815 nodules with surgical pathology results complete CEUS data were retrospectively collected the First People's Hospital Qinzhou between January 2019...

Abstract The computer distance teaching system teaches through the network, and there is no entrance threshold. Any student who willing to study can log in network for at any free time. Neural has a strong self-learning ability an important part of artificial intelligence research. Based on this study, neural network-embedded architecture based shared memory bus structure proposed. By looking alternative method exp function improve speed radial basis algorithm, then by analyzing judgment...

This study aimed to analyze the application of diagnostic model based on deep learning technology in evaluation thyroid contrast-enhanced ultrasound images and provide a reference for benign malignant thyroid. A diagnosis long- short-term memory neural network (LSTM), C-LSTM, was proposed. The method compared with that support vector machine (SVM) manual feature (MF), it applied images. results showed sensitivity, specificity, accuracy C-LSTM were greatly higher than those SVM MF,...

Abstract Based on RFID technology, the tag data can be read and identified dynamically, real-time, contactless, long-distance batch, image information of pointer instrument collected. Using Faster-RCNN to detect locate dial area, combining with Le Net-5 method identify number, system has advantages fast reading, high accuracy, strong adaptability intelligent management. It provides a solution for automatic recognition reading in complex environment veterinary medicine production workshop.

Abstract In this paper, the real-time greenhouse tomato phenotypic parameters data were obtained, and current growth status identified based on in-depth learning method. Combined with under optimal water fertilizer conditions, comparative analysis was carried out. According to difference of parameter tomato, deficiency identified, target scheme determined. Based LSTM neural network model, predicted value environmental compared standard for purpose stopping irrigation fertilization.

Based on the secondary development method of component-based GIS and WaSSI-C model operation mechanism, this paper adopts C / S mode architecture, takes. Net Visual Studio 2012 as environment, use Geodatabase spatial database SQLite relational to store manage running data, realize visualization system by combining C# ArcGIS Engine. The has advantages good human-computer interaction, friendly interface, low cost strong portability. It can not only simulation process based text but also visual...

In this paper, an accurate pre-control system based on veterinary medicine warehouse environment is designed environmental parameter acquisition method, prediction method and device regulation control method. By means of fuzzy C mean clustering denoising, identifying outliers, using cubic spline interpolation to fill the vacant data obtain drug data; The DBSCAN algorithm introduced improve limit learning machine (ELM) predict obtained parameters; adaptive neural network PID regulating...

Abstract In order to realize automatic pruning technology of tomato, an intelligent device tomato lateral branch was designed. A pulley is arranged in the track at top greenhouse planting shed, and a bracket through mechanical connection, positioning module lidar module, secondary close-range camera binocular are middle section crossbar. Through identify locate axil part leaf, arm II fixed main stem plant, judged whether grew leaf terminal visual unit, so as prune or smear inhibitor complete...

In order to realize intelligent plant protection through phenotype in greenhouse, this paper designs a greenhouse system based on discrimination warning by combining mobile device, dispensing barrel, spraying visual device and central control unit. paper, the crop is monitored distinguish type degree of diseases insect pests, early carried out assist administrator select medicament time accurately, automatically medicine, plan navigation route And according position pests real-time...