Real-noise denoising is a challenging task because the statistics of real-noise do not follow the normal distribution, and they are also spatially and temporally changing. In order to cope with various and complex real-noise, we propose a well-generalized denoising architecture and a transfer learning scheme. Specifically, we adopt an adaptive instance normalization to build a denoiser, which can regularize the feature map and prevent the network from overfitting to the training set. We also introduce a transfer learning scheme that transfers knowledge learned from syntheticnoise data to the real-noise denoiser. From the proposed transfer learning, the synthetic-noise denoiser can learn general features from various synthetic-noise data, and the real-noise denoiser can learn the real-noise characteristics from real data. From the experiments, we find that the proposed denoising method has great generalization ability, such that our network trained with synthetic-noise achieves the best performance for Darmstadt Noise Dataset (DND) among the methods from published papers. We can also see that the proposed transfer learning scheme robustly works for real-noise images through the learning with a very small number of labeled data.

BackgroundGel-type autologous chondrocyte (Chondron™) implantations have been used for several years without using periosteum or membrane. This study involves evaluations of the clinical results of Chondron™ at many clinical centers at various time points during the postoperative patient follow-up.MethodsData from 98 patients with articular cartilage injury of the knee joint and who underwent Chondron™ implantation at ten Korean hospitals between January 2005 and November 2008, were included and were divided into two groups based on the patient follow-up period, i.e. 13~24-month follow-up and greater than 25-month follow-up. The telephone Knee Society Score obtained during telephone interviews with patients, was used as the evaluation tool.ResultsOn the tKSS-A (telephone Knee Society Score-A), the score improved from 43.52 ± 20.20 to 89.71 ± 13.69 (P < 0.05), and on the tKSS-B (telephone Knee Society Score-B), the score improved from 50.66 ± 20.05 to 89.38 ± 15.76 (P < 0.05). The total improvement was from 94.18 ± 31.43 to 179.10 ± 24.69 (P < 0.05).ConclusionGel-type autologous chondrocyte implantation for chondral knee defects appears to be a safe and effective method for both decreasing pain and improving knee function.

No abstract

No abstract

Mice lacking a functional Biogenesis of Lysosome-related Organelles Complex 1 (BLOC-1), such as those of the pallid line, display cognitive and behavioural impairments reminiscent of those presented by individuals with intellectual and developmental disabilities. Although disturbances in the sleep/wake cycle are commonly lamented by these individuals, the underlying mechanisms, including the possible role of the circadian timing system, are still unknown. In this paper, we have explored sleep/circadian malfunctions and underlying mechanisms in BLOC-1-deficient pallid mice. These mutants exhibited less sleep behaviour in the beginning of the resting phase than wild-type mice with a more broken sleeping pattern in normal light-dark conditions. Furthermore, the strength of the activity rhythms in the mutants were reduced with significantly more fragmentation and lower precision than in age-matched controls. These symptoms were accompanied by an abnormal preference for the open arm in the elevated plus maze in the day and poor performance in the novel object recognition at night. At the level of the central circadian clock (the suprachiasmatic nucleus, SCN), loss of BLOC-1 caused subtle morphological changes including a larger SCN and increased expression of the relative levels of the clock gene Per2 product during the day but did not affect the neuronal activity rhythms. In the hippocampus, the pallid mice presented with anomalies in the cytoarchitecture of the Dentate Gyrus granule cells, but not in CA1 pyramidal neurones, along with altered PER2 protein levels as well as reduced pCREB/tCREB ratio during the day. Our findings suggest that lack of BLOC-1 in mice disrupts the sleep/wake cycle and performance in behavioural tests associated with specific alterations in cytoarchitecture and protein expression.