ABSTRACT
Unpaired single-image dehazing has become a challenging research hotspot due to its wide application in modern transportation, remote sensing, and intelligent surveillance, among other applications. Recently, CycleGAN-based approaches have been popularly adopted in single-image dehazing as the foundations of unpaired unsupervised training. However, there are still deficiencies with these approaches, such as obvious artificial recovery traces and the distortion of image processing results. This paper proposes a novel enhanced CycleGAN network with an adaptive dark channel prior for unpaired single-image dehazing. First, a Wave-Vit semantic segmentation model is utilized to achieve the adaption of the dark channel prior (DCP) to accurately recover the transmittance and atmospheric light. Then, the scattering coefficient derived from both physical calculations and random sampling means is utilized to optimize the rehazing process. Bridged by the atmospheric scattering model, the dehazing/rehazing cycle branches are successfully combined to form an enhanced CycleGAN framework. Finally, experiments are conducted on reference/no-reference datasets. The proposed model achieved an SSIM of 94.9% and a PSNR of 26.95 on the SOTS-outdoor dataset and obtained an SSIM of 84.71% and a PSNR of 22.72 on the O-HAZE dataset. The proposed model significantly outperforms typical existing algorithms in both objective quantitative evaluation and subjective visual effect.
ABSTRACT
Colorectal cancer (CRC) is one of the main causes of cancer-related deaths worldwide. Sporadic CRC develops from normal mucosa via adenoma to adenocarcinoma, which provides a long screening window for clinical detection. However, early diagnosis of sporadic colorectal adenoma (CRA) and CRC using serum metabolic screening remains unclear. The purpose of this study was to identify some promising signatures for distinguishing the different pathological metabolites of colorectal mucosal malignant transformation. A total of 238 endogenous metabolites were elected. We found that CRA and CRC patients had 72 and 73 different metabolites compared with healthy controls, respectively. There were 20 different metabolites between CRA and CRC patients. The potential metabolites of tumor growth (including patients with CRA and CRC) were found, such as A-d-glucose, D-mannose, N-acetyl-D-glucosamine, L-cystine, Sarcosine, TXB 2, 12-Hete, and chenodeoxycholic acid. Compared with CRA, 3,4,5-trimethoxybenzoic acid was significantly higher in CRC patients. There results prompt us to use the potential serum signatures to screen CRC as the novel strategy. Serum metabolite screening is useful for early detection of mucosal intestinal malignancy. We will further investigate the roles of these promising biomarkers during intestinal tumorigenesis in future. SIGNIFICANCE: CRC is one of the main causes of cancer-related deaths worldwide. Sporadic CRC develops from normal mucosa via adenomas to adenocarcinoma, which provides a long screening window for about 5-10 years. We adopt the metabolic analysis of extensive targeted metabolic technology. The main purpose of the metabolic group analysis is to detect and screen the different metabolites, thereby performing related functional prediction and analysis of the differential metabolites. In our study, 30 samples are selected, divided into 3 groups for metabolic analysis, and 238 metabolites are elected. In 238 metabolites, we find that CRA patients have 72 different metabolites compared with health control. Compared with health control, CRC have 73 different metabolites. Compared with CRA and CRC patients, there are 20 different metabolites. The annotation results of the significantly different metabolites are classified according to the KEGG pathway type. The potential metabolites of tumor growth stage (including patients with CRA and CRC) are found, such as A-d-glucose, D-mannose, N-acetyl-D-glucosamine, L-cystine, sarcosine, TXB 2, 12-Hete and chenodeoxycholic acid. Compared with CRA patients, CRC patients had significantly higher 3,4,5-trimethoxybenzoic acid level. It is prompted to use serum different metabolites to screen CRC to provide new possibilities.
