Voice Transformation Using Two-Level Dynamic Warping and Neural Networks

Voice transformation, for example, from a male speaker to a female speaker, is achieved here using a two-level dynamic warping algorithm in conjunction with an artificial neural network. An outer warping process which temporally aligns blocks of speech (dynamic time warp, DTW) invokes an inner warping process, which spectrally aligns based on magnitude spectra (dynamic frequency warp, DFW). The mapping function produced by inner dynamic frequency warp is used to move spectral information from a source speaker to a target speaker. Artifacts arising from this amplitude spectral mapping are reduced by reconstructing phase information. Information obtained by this process is used to train an artificial neural network to produce spectral warping information based on spectral input data. The performance of the speech mapping compared using Mel-Cepstral Distortion (MCD) with previous voice transformation research, and it is shown to perform better than other methods, based on their reported MCD scores.

Dynamic Warping Network for Semantic Video Segmentation

A major challenge for semantic video segmentation is how to exploit the spatiotemporal information and produce consistent results for a video sequence. Many previous works utilize the precomputed optical flow to warp the feature maps across adjacent frames. However, the imprecise optical flow and the warping operation without any learnable parameters may not achieve accurate feature warping and only bring a slight improvement. In this paper, we propose a novel framework named Dynamic Warping Network (DWNet) to adaptively warp the interframe features for improving the accuracy of warping-based models. Firstly, we design a flow refinement module (FRM) to optimize the precomputed optical flow. Then, we propose a flow-guided convolution (FG-Conv) to achieve the adaptive feature warping based on the refined optical flow. Furthermore, we introduce the temporal consistency loss including the feature consistency loss and prediction consistency loss to explicitly supervise the warped features instead of simple feature propagation and fusion, which guarantees the temporal consistency of video segmentation. Note that our DWNet adopts extra constraints to improve the temporal consistency in the training phase, while no additional calculation and postprocessing are required during inference. Extensive experiments show that our DWNet can achieve consistent improvement over various strong baselines and achieves state-of-the-art accuracy on the Cityscapes and CamVid benchmark datasets.

Modified brute force algorithm to solve the closest pair of points problem based on dynamic warping

<span>This paper introduces an algorithm to solve the closest pair of points problem in a 2D plane based on dynamic warping. The algorithm computes all the distances between the set of points P(x, y) and a reference point R(i, j), records all the result in a grid and finally determines the minimum distance using schematic steps. Results show that the algorithm of finding the closest pair of points has achieved less number of comparisons in determining the closest pair of points compared with the brute force and divide-and-conquer methods of the closest pair of points. </span>

Incremental correlation of multiple well logs following geologically optimal neighbors

Well-log correlation is a crucial step to construct cross sections in estimating structures between wells and building subsurface models. Manually correlating multiple logs can be highly subjective and labor intensive. We have developed a weighted incremental correlation method to efficiently correlate multiple well logs following a geologically optimal path. In this method, we first automatically compute an optimal path that starts with longer logs and follows geologically continuous structures. Then, we use the dynamic warping technique to sequentially correlate the logs following the path. To avoid potential error propagation with the path, we modify the dynamic warping algorithm to use all the previously correlated logs as references to correlate the current log in the path. During the sequential correlations, we compute the geologic distances between the current log and all of the reference logs. Such distances are proportional to Euclidean distances, but they increase dramatically across discontinuous structures such as faults and unconformities that separate the current log from the reference logs. We also compute correlation confidences to provide quantitative quality control of the correlation results. We use the geologic distances and correlation confidences to weight the references in correlating the current log. By using this weighted incremental correlation method, each log is optimally correlated with all the logs that are geologically closer and are ordered with higher priorities in the path. Hundreds of well logs from the Teapot Dome survey demonstrate the efficiency and robustness of the method.