This paper introduces multi-view stereo (MVS) algorithm for an omnidirectional image (ODI). A data and smoothing term for MVS are calculated without distortion of ODIs at high latitudes based on its epipolar geometry. Matching a normal vector with a pole with an ODI removes the distortion on an epipolar curve. Since semi-global matching (SGM) is a one-dimensional smoothing method, a combination of SGM can wholly remove from the distortion of the ODI. As a result of experiments, disparity and depth can be calculated using the proposed method.

To create realistic images of outdoor scenes using computer graphics techniques, synthetic clouds with high-resolution details are required. However, the computation cost for generating high-resolution clouds is very expensive and time-consuming. In this paper, we address this problem and propose an efficient method to create high-resolution images of clouds. The method combines low resolution 3D simulation with high resolution 2D simulation. The low resolution 3D simulation is fast and is used for determining the overall shape of the clouds. Then the high resolution 2D simulation uses the result of the 3D simulation and adds small-scale details to them. We demonstrate that our method achieves four to ten times faster computation to generate realistic high-resolution images of clouds.

To date, there is a wide range of registration methods for 3D point clouds. Nevertheless, most approaches are inadequate for pairs of 3D point clouds with low overlapping ratios due to many incorrect estimates of putative point correspondences when numerous outliers are present. In this particular scenario, points from the non-overlapping areas represent the outliers. In this work, we introduce a piecewise voting-based method that leverages the repeatability of the transformation matrices that result from registering point subsets of the overlapping areas. We show the performance of our approach under the effect of its ruling parameters through extensive ablation experiments to define an easy-to-follow usage guideline and compare it to available state-of-the-art registration methods for low overlap 3D points clouds.

The use of video clip sharing sites and applications has become popular in the recent years. However, when video contributors were looking for sound sources that could be used as background music, it took a lot of time to actually listen to a number of songs to confirm if they fit the video. The author of this study believes that video contributors would improve their production efficiency if there is a system that automatically lists up suitable background music for videos. In the proposed method, the features for the videos are formulated using color and optical flow histograms. Mel frequency cepstrum coefficients, which have often been used in the recent deep learning research, are used for the music features. A function that converts these features to linguistic evaluation values is obtained by multiple regression analysis. The top five music files with the highest similarity to the input video are calculated to evaluate the proposed system. Whether they are compatible with the respondents’ ranking is then checked. The result shows that the proposed method via impression words can perform a more accurate retrieval than the learning network that directly connects video and music features.

This paper presents a simple approach to encode video sequences by selecting an optimal Intra encoded frame (I-frame). I-frames in Versatile Video Coding (VVC) random access mode influence the bitrate required for the remaining frames in the group of pictures (GOP). Hence, we carry out a full search for the optimal frame in the first GOP and select an I-frame with appropriate spatiotemporal information. This full search is carried out during the pre-processing stage and hence, do not introduce large complexities into the VVC codec. We prove our technique with over 100 sequences, using BVI-DVC and JVET sequences, to have wide range of content. Our technique is best suited for offline coding and suitable for pre-recorded streaming videos. Applying this technique to encode JVET sequences using the VVC random access mode, results indicate average bit rate savings (based on the Bjøntegaard Delta metric) of approximately 2.1% on the Y-channel.