Due to the absorption and scattering effects of light in water bodies and the non-uniformity and insufficiency of artificial illumination, underwater images often present various degradation problems, impacting their utility in underwater applications. In this paper, we propose a model-based underwater image simulation and learning-based underwater image enhancement method for coping with various degradation problems in underwater images. We first derive a simplified model for describing various degradation problems in underwater images, then propose a model-based image simulation method that can generate images with a wide range of parameter values. The proposed image simulation method also comes with an image-selection part, which helps to prune the simulation dataset so that it can serve as a training set for learning to enhance the targeted underwater images. Afterwards, we propose a convolutional neural network based on the encoder-decoder backbone to learn to enhance various underwater images from the simulated images. Experiments on simulated and real underwater images with different degradation problems demonstrate the effectiveness of the proposed underwater image simulation and enhancement method, and reveal the advantages of the proposed method in comparison with many state-of-the-art methods.

Background: Increased level of serum macrophage inhibitory cytokine-1 (MIC-1), a member of transforming growth factor-μ superfamily, was found in patients with epithelial tumors. This study aimed to evaluate whether serum level of MIC-1 can be a candidate diagnostic and prognostic indicator for early-stage nonsmall cell lung cancer (NSCLC). Methods: A prospective study enrolled 152 patients with Stage I–II NSCLC, who were followed up after surgical resection. Forty-eight patients with benign pulmonary disease (BPD) and 105 healthy controls were also included in the study. Serum MIC-1 levels were measured using an enzyme-linked immunosorbent assay, and the association with clinical and prognostic features was analyzed. Results: In patients with NSCLC, serum protein levels of MIC-1 were significantly increased compared with healthy controls and BPD patients (all P< 0.001). A threshold of 1000 pg/ml of MIC-1 was found in patients with early-stage (Stage I and II) NSCLC, with sensitivity and specificity of 70.4% and 99.0%, respectively. The serum levels of MIC-1 were associated with age (P = 0.001), gender (P = 0.030), and T stage (P = 0.022). Serum MIC-1 threshold of 1465 pg/ml was found in patients with poor early outcome, with sensitivity and specificity of 72.2% and 66.1%, respectively. The overall 3-year survival rate of NSCLC patients with high serum levels of MIC-1 (≥1465 pg/ml) was lower than that of NSCLC patients with low serum MIC-1 levels (77.6% vs. 94.8%). Multivariate Cox regression survival analysis showed that a high serum level of MIC-1 was an independent risk factor for reduced overall survival (hazard ratio = 3.37, 95% confidential interval: 1.09–10.42, P= 0.035). Conclusion: The present study suggested that serum MIC-1 may be a potential diagnostic and prognostic biomarker for patients with early-stage NSCLC.

Background: In order to minimize the injury reaction during the surgery and reduce the morbidity rate, hence reducing the mortality rate of esophagectomy, minimally invasive esophagectomy (MIE) was introduced. The aim of this study was to compare the postoperative outcomes in patients with esophageal squamous cell carcinoma undergoing minimally invasive or open esophagectomy (OE). Methods: The medical records of 176 consecutive patients, who underwent minimally invasive esophagectomy (MIE) between January 2009 and August 2013 in Cancer Institute & Hospital, Chinese Academy of Medical Sciences, were retrospectively reviewed. In the same period, 142 patients who underwent OE, either Ivor Lewis or McKeown approach, were selected randomly as controls. The clinical variables of paired groups were compared, including age, sex, Charlson score, tumor location, duration of surgery, number of harvested lymph nodes, morbidity rate, the rate of leak, pulmonary morbidity rate, mortality rate, and hospital length of stay (LOS). Results: The number of harvested lymph nodes was not significantly different between MIE group and OE group (median 20 vs. 16, P=0.740). However, patients who underwent MIE had longer operation time than the OE group (375 vs. 300 minutes, P <0.001). Overall morbidity, pulmonary morbidity, the rate of leak, in-hospital death, and hospital LOS were not significantly different between MIE and OE groups. Morbidities including anastomotic leak and pulmonary morbidity, in-hospital death, hospital LOS, and hospital expenses were not significantly different between MIE and OE groups as well. Conclusions: MIE and OE appear equivalent with regard to early oncological outcomes. There is a trend that hospital LOS and hospital expenses are reduced in the MIE group than the OE group.

Comparative study of video-assisted thoracic surgery versus open thymectomy for thymoma in one single center

To meet the changing real-time edge engineering application requirements of CNN, aiming at the lack of universality and flexibility of CNN hardware acceleration architecture based on ARM+FPGA, a general low-power all pipelined CNN hardware acceleration architecture is proposed to cope with the continuously updated CNN algorithm and accelerate in hardware platforms with different resource constraints. In the framework of the general hardware architecture, a basic instruction set belonging to the architecture is proposed, which can be used to calculate and configure different versions of CNN algorithms. Based on the instruction set, the configurable computing subsystem, memory management subsystem, on-chip cache subsystem, and instruction execution subsystem are designed and implemented. In addition, in the processing of convolution results, the on-chip storage unit is used to preprocess the convolution results, to speed up the activation and pooling calculation process in parallel. Finally, the accelerator is modeled at the RTL level and deployed on the XC7Z100 heterogeneous device. The lightweight networks YOLOv2-tiny and YOLOv3-tiny commonly used in engineering applications are verified on the accelerator. The results show that the peak performance of the accelerator reaches 198.37 GOP/s, the clock frequency reaches 210 MHz, and the power consumption is 4.52 w under 16-bit width.

Underwater wireless optical communication (UWOC) is acknowledged as a useful way to transmit data in the ocean for short-distance applications. Carrying a UWOC device on mobile platforms is quite practical in ocean engineering, which is helpful to exploit its advantages. In application, such a platform needs a camera to observe the surroundings and guide its action. Since the majority of ocean is always dark, active illumination is necessary to imaging. When UWOC works in such an environment, its performance is affected by the illumination light noise. In this paper, we study the influence of underwater LED illumination on bidirectional UWOC with the Monte Carlo method. We simulate forward noise from LED illumination to the opposite receiver in the cooperative terminal, and the backscattering noise on the adjacent receiver in the same terminal. The results show that the forward noise is reduced with the increase of theabsorption coefficient, scattering coefficient, transmitting distance, and separated distance between receiver and the optical axis of LED. However, it becomes greater with the field of view (FOV) of the receiver. The backscattering noise is reduced with the increase of the absorption coefficient and separated distance between receiver and LED. However, it becomes greater with the FOV and scattering coefficient, while it has little relation with transmitting distance. In order to reduce these two kinds of noises, besides inserting an optical filter in the receivers and narrowing their FOV, the optical axis of LED light should keep away from the receivers. The results in this paper are helpful for UWOC application.