Editorial Issue 34.2

Abstract

This issue contains 10 papers. In the first paper, Diego Mazala and Claudio Esperança, from Universidade Federal do Rio de Janeiro in Brazil, and Ricardo Marroquim, from Delft University of Technology in Netherlands, present a novel method for blending human faces in order to create a new one. In a nutshell, their proposal uses Laplacian smoothing to segregate layers of details from one or more faces, which are then integrated into a base face with the help of an interactive and visual editor. In particular, their method supports blending multiple faces and multiple subregions in those faces. Since their approach is intuitive and relatively easy to implement, it can be integrated into artistic pipelines aiming at designing human face models from preexisting ones. In the second paper, Ulas Gulec, from TED University Ankara, Turkey, Ilkin Sevgi Isler, from University of Central Florida in Orlando, USA, Mehmet Harun Doganay and Muruvet Gokcen, from Milsoft Software Technologies, in Ankara, Turkey, Mert Ali Gozcu, from Simsoft Information Technologies in Ankara, Turkey, and Merve Denizci Nazligul, from Yeditepe University in Istanbul, Turkey, propose a study that aims to increase the motivation levels of powerlifters during training sessions by developing a virtual competition environment. In this environment, the athletes experience a virtual competition environment by using HTC Vive. To understand the efficiency of the virtual environment, it was tested with 32 professional athletes. The findings illustrated that using VR technology was beneficial to increase the level of motivation of powerlifters during training sessions. In the third paper, Jinyu Li, Xin Zhou, Bangbang Yang, Guofeng Zhang, Xun Wang, and Hujun Bao, from Zhejiang University in Hangzhou, China, propose RLP-VIO—a robust and lightweight monocular visual-inertial odometry system using multiplane priors. With planes extracted from the point cloud, visual-inertial-plane PnP uses the plane information for fast localization. For sensor fusion, their sliding-window optimization uses a novel structureless plane-distance error cost, which prevents the fill-in effect that poisons the BA problem's sparsity and permits the use of a smaller sliding window while maintaining good accuracy. The total computational cost is further reduced with their modified marginalization strategy. To further improve the tracking robustness, the landmark depths are constrained using the planes during degenerated motion. The whole system is parallelized with a three-stage pipeline. Their system achieves competitive accuracy and works robustly even on long and challenging sequences. In the fourth paper, Ruizhe Li, Ryo Oji, and Issei Fujishiro, from Keio University in Yokohama, Japan, present an anime-like character customization system, where each customizing parameter can adjust the shape or color for the corresponding part of the character model. Based on this system, the authors propose an improved approach for generating a rich variety of 3D anime-like NPCs including body models and clothing items in different styles. They introduce a neural network to control the facial appearances, Gaussian mixture models to control the colors of hair and clothes, and a Bayesian network to control the outfits of clothing items. They demonstrate the proposed approach can maintain variety and stability for generated characters. The fifth paper by Cheng Shang, Hongke Zhao, Meili Wang, Xiao Long Wang, Yu Jiang, and Qiang Gao, from Northwest Agriculture and Forestry University in Yangling, China, focuses on the identification of cashmere goats with similar characteristics. First, the single shot detection network was used to process the data set. Next, the authors innovatively proposed the multibranch fusion optimization structure of triplet loss function and Label Smoothing CrossEntropy Loss function, as well as they added a small number of images of 24 different breeds of sheep to each cashmere goat dataset with different ID to promote the distance between training individuals, and then used the trained model to find the number of goats with the lowest recognition accuracy. Unlike previous studies using the Cycle-GAN, the authors took the novel approach of using this network to learn and combine the features seen in photos of cashmere goats. Since the learned features were all observed in the same goats, this method achieved better results in learning the features of the goats. In the sixth paper, Jia Chen, Haidongqing Yuan, Yi Zhang, Ruhan He, and Jinxing Liang, from Wuhan Textile University in China, propose a fashion image retrieval framework based on dilated convolutional residual network that consists of two major parts, image feature extraction and feature distance measurement. For image feature extraction, they first extract the shallow features of the input image by a multiscale convolutional network, and then develop a novel dilated convolutional residual network to obtain the deep features of the image. Finally, the extracted features are transformed into high-dimensional features vector by a binary retrieval vector module. For feature distance measurement, the authors first use PCA to reduce the dimension of the extracted high-dimensional vectors. Then they propose a mixed distance measurement algorithm combined with cosine distance and Mahalanobis distance to calculate the spatial distance of the feature vectors for similarity ranking. In the seventh paper, Yi-Jheng Huang, from Yuan Ze University in Taoyuan, Taiwan, proposes an algorithm for detecting edges based on the color of a mesh surface. His approach is based on the data structure of a quad mesh, which makes the data structure of 3D meshes resemble the data structure of images. As a result, image-processing methods can be applied on the 3D meshes. In this paper, six classical edge detection filters are implemented on the 3D meshes. The experimental results demonstrate that his method can identify areas of high color gradient on 3D meshes. A comparison with two other methods for detecting color boundaries on 3D meshes reveals that his method is more effective at detecting boundaries. Lastly, he proposes two novel applications that utilize the information of color boundaries on a 3D mesh surface. In the eighth paper, Xue Du, Juan Xiu Sun, Kunpeng Wang, Junlong Yang, and Jiang Chuan Wang, from Shandong University of Science and Technology in China, propose an underwater image enhancement method based on entropy weight fusion for underwater images. First, white balance processing can effectively correct the blue (green) color appearance of the image. Then the white-balanced images are converted from RGB space to LAB space, and L channel is processed with improved adaptive gamma correction, and then converted back to RGB space. CLAHE and bilateral filtering are performed in RGB space. The RGB space is converted to HSV space, the V channel is processed by single-scale Retinex algorithm combined with guided filtering, stretching the R channel with s-cosine curve, and then converted back to RGB space. Finally, the three results are fused by entropy weight to obtain the final enhanced image. Experimental results show that the proposed algorithm can improve the contrast and clarity of underwater images, and effectively remove color cast. In the ninth paper, Ana Agić, Lidija Mandić, and Lea Skorin-Kapov, from University of Zagreb in Croatia, report on the results of a user study aimed to investigate and compare three types of locomotion techniques in VR in terms of their impact on cybersickness. The different locomotion techniques are tested in two different contrast scene settings, daytime and nighttime, to further explore the potential impact of scene contrast adjustments. For the evaluation of cybersickness, the authors used a questionnaire to obtain subjective ratings, and heart rate monitoring as an objective metric. Results show that a linear movement locomotion technique provokes the highest level of cybersickness, and that women have a higher heart rate as compared with men when navigating and interacting in a VR scene. Regarding the influence of scene contrast, results showed that scenes with daylight were better suited to participants in almost all tested scenarios. In addition to reported findings related to locomotion techniques and the impact on cybersickness, they highlight that a key contribution is the utilized test methodology. The last paper by Zafar Masood, Jiangbin Zheng, Muhammad Irfan, and Idrees Ahmad, from Northwestern Polytechnical University in Xi'an, China, presents a novel method for a high-performance large-scale terrain rendering for high-fidelity display systems using game engine. The proposed method performs patch-based hierarchical culling of a multiresolution terrain model to reduce rendering load. A view-based algorithm simplifies the patches with error control on GPU. Simplified patches are efficiently submitted for drawing using indirect mesh instancing feature of game engine. The method utilizes hardware tessellation feature for high-performance model tessellation and accurate Earth's surface construction using displacement mapping. The proposed method is evaluated by rendering scenes for high-quality output on consumer-level hardware. Flights are performed with various settings and results are compared with clipmap-based and state-of-the-art hardware tessellation based adaptive methods. The method achieved 750, 575, and 540 frames-per-second (fps) for HD, full-HD, and ultra-HD display resolutions.