Generative Image Inpainting with Dilated Deformable Convolution

The image generation and completion model complement the missing area of the image to be repaired according to the image itself or the information of the image library so that the repaired image looks very natural and difficult to distinguish from the undamaged image. The difficulty of image generation and completion lies in the reasonableness of image semantics and the clear and true texture of the generated image. In this paper, a Wasserstein generative adversarial network with dilated convolution and deformable convolution (DDC-WGAN) is proposed for image completion. A deformable offset is added based on dilated convolution, which enlarges the receptive field and provides a more stable representation of geometric deformation. Experiments show that the DDC-WGAN method proposed in this paper has better performance in image generation and complementation than the traditional generative adversarial complementation network.

Multi-Modal Remote Sensing Image Matching Method Based on Deep Learning Technology

Remote sensing is a scientific technology that uses sensors to detect the reflection, radiation or scattering of electromagnetic wave signals from ground objects in a non-contact and long-distance manner. The images are classified by the extracted image feature information Recognition is a further study of obtaining target feature information, which is of great significance to urban planning, disaster monitoring, and ecological environment evaluation. The image matching framework proposed in this paper matches the depth feature maps, and reversely pushes the geometric deformation between the depth feature maps to between the original reference image and the target image, and eliminates the geometric deformation between the original images. Finally, through feature extraction of the corrected image, the extracted local feature image blocks are input into the trained multi-modal feature matching network to complete the entire matching process. Experiments show that the negative sample set construction strategy that takes into account the sample distance proposed in this experiment can effectively deal with the problem of neighboring point interference in RSI matching, and improve the matching performance of the network model.

Luminescence Properties of Tetrahedral Coordinated Mn2+; Genthelvite and Willemite Examples

The cause of the split of 4A4E(4G) Mn2+ excited level measured on minerals spectra is discussed. It is our view that ∆E = |4E(4G) − 4A(4G)| should be considered an important spectroscopic parameter. Among the possible reasons for the energy levels splitting taken under consideration, such as the covalent bond theory, the geometric deformation of the coordination polyhedron and the lattice site’s symmetry, the first one was found to be inappropriate. Two studied willemite samples showed that the impurities occur in one of the two available lattice sites differently in both crystals. Moreover, it was revealed that the calculated crystal field Dq parameter can indicate which of the two non-equivalent lattice sites positions in the willemite crystal structure was occupied by Mn2+. The above conclusions were confirmed by X-ray structure measurements. Significant differences were also noted in the Raman spectra of these willemites.

An improved YOLOv3 model for detecting location information of ovarian cancer from CT images

Ovarian cancer is a malignant tumor that poses a serious threat to women’s lives. Computer-aided diagnosis (CAD) systems can classify the type of ovarian tumors, but few of them can provide exactly the location information of ovarian cancer cells. Recently, deep learning technology becomes hot for automatic detection of cancer cells, particularly for detecting their locations. In this work, we propose a novel end-to-end network YOLO-OC (Ovarian cancer) model, which can extract the characteristics of ovarian cancer more efficiently. In our method, deformable convolution is used to enhance the model’s ability to learn geometric deformation in space. Squeeze-and-Excitation (SE) module is proposed to automatically learn the importance of different channel features. Data experiments are conducted on datasets collected from The Affiliated Hospital of Qingdao University Medical College, China. Experimental results show that our YOLO-OC model achieves 91.83%, 85.66% and 73.82% on mean average precision [email protected], [email protected] and mAP@[.5,.95], respectively, which performs better than Faster R-CNN, SSD and RetinaNet on both accuracy and efficiency.

Minimally deformed anisotropic stars by gravitational decoupling in Einstein–Gauss–Bonnet gravity

In this article, we develop a theoretical framework to study compact stars in Einstein gravity with the Gauss–Bonnet (GB) combination of quadratic curvature terms. We mainly analyzed the dependence of the physical properties of these compact stars on the Gauss–Bonnet coupling strength. This work is motivated by the relations that appear in the framework of the minimal geometric deformation approach to gravitational decoupling (MGD-decoupling), we establish an exact anisotropic version of the interior solution in Einstein–Gauss–Bonnet gravity. In fact, we specify a particular form for gravitational potentials in the MGD approach that helps us to determine the decoupling sector completely and ensure regularity in interior space-time. The interior solutions have been (smoothly) joined with the Boulware–Deser exterior solution for 5D space-time. In particular, two different solutions have been reported which comply with the physically acceptable criteria: one is the mimic constraint for the pressure and the other approach is the mimic constraint for density. We present our solution both analytically and graphically in detail.

Moving Car Recognition and Removal for 3D Urban Modelling Using Oblique Images

With the progress of photogrammetry and computer vision technology, three-dimensional (3D) reconstruction using aerial oblique images has been widely applied in urban modelling and smart city applications. However, state-of-the-art image-based automatic 3D reconstruction methods cannot effectively handle the unavoidable geometric deformation and incorrect texture mapping problems caused by moving cars in a city. This paper proposes a method to address this situation and prevent the influence of moving cars on 3D modelling by recognizing moving cars and combining the recognition results with a photogrammetric 3D modelling procedure. Through car detection using a deep learning method and multiview geometry constraints, we can analyse the state of a car’s movement and apply a proper preprocessing method to the geometrically model generation and texture mapping steps of 3D reconstruction pipelines. First, we apply the traditional Mask R-CNN object detection method to detect cars from oblique images. Then, a detected car and its corresponding image patch calculated by the geometry constraints in the other view images are used to identify the moving state of the car. Finally, the geometry and texture information corresponding to the moving car will be processed according to its moving state. Experiments on three different urban datasets demonstrate that the proposed method is effective in recognizing and removing moving cars and can repair the geometric deformation and error texture mapping problems caused by moving cars. In addition, the methods proposed in this paper can be applied to eliminate other moving objects in 3D modelling applications.

Ultracompact stars with polynomial complexity by gravitational decoupling

In this work we construct an ultracompact star configuration in the framework of Gravitational Decoupling by the Minimal Geometric Deformation approach. We use the complexity factor as a complementary condition to close the system of differential equations. It is shown that for a polynomial complexity the resulting solution can be matched with two different modified-vacuum geometries.