https://www.selleckchem.com/ Clinical abdominal CT angiographic images of 200 patients were applied to perform the evaluation. RESULTS Extensive experimental results show that the proposed method achieves a higher dice coefficient (DSC) of 0.826 than the other two existing weakly-supervised deep neural networks. Furthermore, the segmentation performance is close to the fully supervised deep CNN. CONCLUSIONS The proposed strategy improves not only the efficiency of network training but also the precision of the segmentation.BACKGROUND Clostridium difficile infection (CDI), especially hospital-acquired Clostridium difficile infection (HA-CDI), continues to be a public health problem and has aroused great concern worldwide for years. This study aimed to elucidate the clinical and epidemiological features of HA-CDI and the characteristics of C.difficile isolates in Chongqing, Southwest China. METHODS A case-control study was performed to identify the clinical incidence and risk factors of HA-CDI. C. difficile isolates were characterised by polymerase chain reaction (PCR) ribotyping, multilocus sequence typing (MLST), toxin gene detection and antimicrobial susceptibility testing. RESULTS Of the 175 suspicious patients, a total of 122 patients with antibiotic-associated diarrhea (AAD) were included in the study; among them, 38 had HA-CDI. The incidence of AAD and HA-CDI was 0.58 and 0.18 per 1000 patient admissions, respectively. Chronic renal disease and cephalosporin use were independent risk factors for HA-CDI. Fifty-five strains were assigned into 16 sequence types (STs) and 15 ribotypes (RTs). ST2/RT449 (8, 14.5%) was the predominant genotype. Of the 38 toxigenic isolates, A + B + CDT- isolates accounted for most (34, 89.5%) and 1 A + B + CDT+ isolate emerged. No isolate was resistant to vancomycin, metronidazole or tigecycline, with A-B-CDT- being more resistant than A + B + CDT-. CONCLUSIONS Different genotypes of C. difficile strains were witnessed in Chongqing, w