RESUMO
Enhancing the visibility in extreme low-light environments is a challenging task. Under nearly lightless condition, existing image denoising methods could easily break down due to significantly low SNR. In this paper, we systematically study the noise statistics in the imaging pipeline of CMOS photosensors, and formulate a comprehensive noise model that can accurately characterize the real noise structures. Our novel model considers the noise sources caused by digital camera electronics which are largely overlooked by existing methods yet have significant influence on raw measurement in the dark. It provides a way to decouple the intricate noise structure into different statistical distributions with physical interpretations. Moreover, our noise model can be used to synthesize realistic training data for learning-based low-light denoising algorithms. In this regard, although promising results have been shown recently with deep convolutional neural networks, the success heavily depends on abundant noisy-clean image pairs for training, which are tremendously difficult to obtain in practice. Generalizing their trained models to images from new devices is also problematic. Extensive experiments on multiple low-light denoising datasets - including a newly collected one in this work covering various devices - show that a deep neural network trained with our proposed noise formation model can reach surprisingly-high accuracy. The results are on par with or sometimes even outperform training with paired real data, opening a new door to real-world extreme low-light photography.
RESUMO
In this article, we propose an alternating directional 3-D quasi-recurrent neural network for hyperspectral image (HSI) denoising, which can effectively embed the domain knowledge-structural spatiospectral correlation and global correlation along spectrum (GCS). Specifically, 3-D convolution is utilized to extract structural spatiospectral correlation in an HSI, while a quasi-recurrent pooling function is employed to capture the GCS. Moreover, the alternating directional structure is introduced to eliminate the causal dependence with no additional computation cost. The proposed model is capable of modeling spatiospectral dependence while preserving the flexibility toward HSIs with an arbitrary number of bands. Extensive experiments on HSI denoising demonstrate significant improvement over the state-of-the-art under various noise settings, in terms of both restoration accuracy and computation time. Our code is available at https://github.com/Vandermode/QRNN3D.
RESUMO
OBJECTIVE: To investigate changes in the upper airway and its surrounding soft tissue and to characterize the extent and severity of upper airway obstruction in 136 obstructive sleep apnea/hypopnea syndrome (OSAHS) patients who were awake. METHODS: OSAHS patients and healthy controls were evaluated by fiber-optic nasolaryngoscopy and MRI. The cross-sectional area and pharyngeal wall thickness of the retropalatal, retroglossal and epiglottic region were determined. RESULTS: Ninety-five percent of the mild OSAHS subjects had single-plane obstruction (vs. severe OSAHS, p < 0.05), 5.0% of the mild OSAHS subjects had two-plane obstruction (p < 0.05) and none of them had three-plane obstruction (p < 0.05). The cross-sectional area of the retropalatal, retroglossal and epiglottic region progressively declined as the severity of OSAHS increased (severe OSAHS vs. controls, p < 0.05). The lateral pharyngeal wall was significantly thicker in OSAHS subjects than in healthy controls (p < 0.05). The cross-sectional area of the soft palate in moderate and severe OSAHS subjects was markedly larger than that of the healthy controls (p < 0.05) and positively correlated with the apnea/hypopnea index (p < 0.05). CONCLUSION: Moderate and severe OSAHS patients exhibit multi-plane obstruction of the upper airway, particularly in the retropalatal and retroglossal region. The severity of OSAHS negatively correlates with the thickness, length and cross-sectional area size of the soft palate.
