The ocean connects all continents and is an important space for human activities. Ship detection with electro-optical images has shown great potential due to the abundant imaging spectrum and, hence, strongly supports human activities in the ocean. A suitable imaging spectrum can obtain effective images in complex marine environments, which is the premise of ship detection. This paper provides an overview of ship detection methods with electro-optical images in marine environments. Ship detection methods with sea–sky backgrounds include traditional and deep learning methods. Traditional ship detection methods comprise the following steps: preprocessing, sea–sky line (SSL) detection, region of interest (ROI) extraction, and identification. The use of deep learning is promising in ship detection; however, it requires a large amount of labeled data to build a robust model, and its targeted optimization for ship detection in marine environments is not sufficient.

Stereo vision can be achieved with a single camera aided by a flat-face prism. However, large optical distortions may be caused by irregular refraction. To mitigate this deficiency and minimize distortion, a new freeform prism design is proposed to replace the flat-face prism. All refracting surfaces of the prism are designed for freeform features, which can process object points at a large depth of field. Differential equations are derived to express the freeform surface, and the Runge-Kutta method is used to numerically solve the differential equations.

Unilateral breast reconstruction is different from bilateral breast reconstruction and breast augmentation, as it requires more consideration of the symmetry between the reconstructed breast and the contralateral breast. However, there are currently no implants specifically targeted to unilateral breast reconstruction, nor is there is objective classification criteria for breast appearance type. A total of 212 breasts from 153 patients were measured. Breast measurements, breast surface curvature distributions, and elliptic Fourier coefficients were used to describe breast features. We objectively classified the appearance of breasts using machine learning and used the resulting cluster centers to construct the most adaptive implants for each type of breast. All of the breasts clustered into 4 types. The implants corresponding to each type of breast were constructed using cluster centers. The resulting cluster centers were then used to choose a suitable implant for patients requiring unilateral breast reconstruction. Contour coefficients were used to evaluate the clustering results, with an average score of 0.53. The two breasts that develop normally in the same person were treated as the same class. The score obtained after statistical classification was 0.47. These results demonstrate that our proposed method can improve the classification of breasts of different shapes. This method provides a foundation for improving the symmetry of unilateral breast reconstruction and the construction and selection of implants.

This paper proposes a deep convolutional neural network-based low-light image enhancement method. In order to adaptively enhance the image brightness, a convolutional neural network with convolutional modules is designed. Lowlight image is firstly down-sampled into sub-images. Then an illumination map is obtained from the input image to provide additional information to the network. The network works on a tensor that consists of sub-images and illumination map, achieving a good performance in brightness increasing and structure preservation. The enhanced result is reconstructed from the output sub-images. Experimental results demonstrate the effectiveness of the proposed method in low-light image enhancement.

In this paper, we consider the construction of irregular QC-LDPC codes by jointly optimizing the girth, the number and approximate cycle extrinsic message degree (ACE) of short cycles to further improve the error-floor performance of irregular QC-LDPC codes with given degree distributions. By selecting the best edges, each of which must make the cycles consisting of the current variable node have the best lower bound of ACE or the minimum number of cycles with ACE lower bound, for each variable node via the progressive Edge growth (PEG) approaches, irregular QC-LDPC codes with better tradeoff between girth, the shortest cycle number and ACE could be constructed. Simulation results show that the proposed codes achieve better error-floor performance compared with the codes using the same degree distributions constructed by the recent proposed methods.