How to construct low-altitude aerial image datasets for deep learning.

The combination of Unmanned Aerial Vehicle (UAV) technologies and computer vision makes UAV applications more and more popular. Computer vision tasks based on deep learning usually require a large amount of task-related data to train algorithms for specific tasks. Since the commonly used datasets are not designed for specific scenarios, in order to give UAVs stronger computer vision capabilities, large enough aerial image datasets are needed to be collected to meet the training requirements. In this paper, we take low-altitude aerial image object detection as an example to propose a framework to demonstrate how to construct datasets for specific tasks. Firstly, we introduce the existing low-altitude aerial images datasets and analyze the characteristics of low-altitude aerial images. On this basis, we put forward some suggestions on data collection of low-altitude aerial images. Then, we recommend several commonly used image annotation tools and crowdsourcing platforms for data annotation to generate labeled data for model training. In addition, in order to make up the shortage of data, we introduce data augmentation techniques, including traditional data augmentation and data augmentation based on oversampling and generative adversarial networks.

Analysis of ROC: The value of HPV16 E6 protein in the diagnosis of early stage cervical carcinoma and precancerous lesions.

Cervical carcinoma is a multifactorial malignant tumor and diagnosis is therefore crucial. The aim of the present study was to examine the value of E6 oncoprotein, in human papillomavirus type 16 (HPV16), in the diagnosis of early stage cervical carcinoma and precancerous lesions. Receiver operating characteristic curve was used to analyze accuracy of diagnosis. A total of 124 patients infected with HPV16 were included in the study. The patients had an average age of 46.7±6.9 years and duration of disease of 10.5±3.4 months. To determine the expression level of HPV16 E6 the immunohistochemical Elivision method was performed. Proportion/horizon positive cells were used to count the cells, and pathologic diagnosis was employed for analysis of the results. The average follow-up time was 2.6±0.7 years. Sensitivity and specificity of diagnosing HPV16 E16 at 1 and 2 years, respectively, were calculated. The diagnostic rate of cervical carcinoma increased with time, and the positive expression of HPV16 E6 was also increased with the development of the disease. Differences among groups were statistically significant (P<0.05). Sensitivity, specificity and accuracy (AUC) of HPV16 E6 diagnosis improved with time, and the differences were statistically significant (P<0.05). Thus, HPV16 E6 oncoprotein can be used as an indicator with good sensitivity and specificity to diagnose early cervical carcinoma and precancerous lesions. The results therefore showed that accuracy increased with the development of the disease.

Transvenous treatment of complex cavernous dural arteriovenous fistulae with Onyx and coils.

Transvenous embolization has become the treatment of choice for such lesions We evaluated Onxy for patients with cavernous dural arteriovenous fistulae (CDAVFs) who underwent transvenous embolization via different transvenous approaches. Case records of six patients with symptomatic CDAVFs, treated between October 2006 and November 2007 were reviewed. A total of seven transvenous procedures were performed in the six patients with CDAVFs. All the patients with CDAVFs of the cavernous sinus were symptom free following embolization. The approach via the internal jugular vein and the inferior petrosal sinus was possible in four of the six patients, with complete occlusion of the fistula. In the remaining two patients, the approach was via the facial vein. Transient bradyarrythmia without morbidity was the only complication in two patients.

Detection of drug resistance-associated genes of multidrug-resistant Acinetobacter baumannii.

The beta-lactamase (BLA) genes, the genes for aminoglycosides-modifying enzymes (AMEs), disinfectant-sulfanilamide resistance (qacEDelta1-sul1) genes, class 1 integrase (intl1) gene, and the qnr gene associated with plasmid-mediated quinolone resistance were analyzed using PCR and verified by DNA sequencing for 31 clinical isolates of multidrug-resistant Acinetobacter baumannii (MDRAB). The organism typing was performed by pulsed-field gel electrophoresis (PFGE). The positive rate of ADC, TEM, PER, and DHA of BLA genes were 100%, 61.3%, 19.4%, and 3.2%, respectively; however, the genes of SHV, OXA-23 group, OXA-24 group, GES, VIM, IMP, and qnr gene were negative. The positive rate of the genes of AMEs for aac (3)-I, aac (6')-I, ant (3")-I, ant (2")-I, aac (3)-II, and aac (6')-II were 67.7%, 45.2%, 29.0%, 22.6%, 12.9%, and 3.2%, respectively. The positive rate of qacEDelta1-sul1 and intl1 were 80.6% and 58.1%, respectively. Six different PFGE clones were found, of which two dominated. The findings show that clinical isolates of MDRAB harbor various kinds of resistance genes.