Dynamic time warping

About: Dynamic time warping is a research topic. Over the lifetime, 6013 publications have been published within this topic receiving 133130 citations.

A Robust Step Detection and Stride Length Estimation for Pedestrian Dead Reckoning Using a Smartphone

Abstract: As an infrastructure-free positioning and navigation method, pedestrian dead reckoning (PDR) is still a research hotspot in the field of indoor localization. Step detection (SD) and stride length estimation (SLE) are two key components of PDR, and it is a challenging problem to apply SD and SLE to different walking patterns. Focusing on this problem, this paper proposes a robust SD and SLE method based on recognizing three walking patterns (i.e., Normal Walk, March in Place, and Quick Walk) using a smartphone. First, we propose a dynamic time warping–based peak prediction with zero-crossing detection to improve the SD accuracy. In particular, the proposed SD can accurately identify the starting and ending points of each step in the three walking patterns. Second, according to the extracted features of each step, a random forest algorithm with classification proofreading is used to recognize the three walking patterns. Finally, an improved SLE model is proposed for the different walking patterns to achieve a higher SLE accuracy. The experimental results show that, on average, the SD accuracy is about 97.9%, the recognition accuracy is about 98.4%, and the relative error of the estimated walking distance is about 3.0%, which outperforms those of the existing commonly used SD and SLE methods.

Time series classification: nearest neighbor versus deep learning models

Abstract: Time series classification has been an important and challenging research task. In different domains, time series show different patterns, which makes it difficult to design a global optimal solution and requires a comprehensive evaluation of different classifiers across multiple datasets. With the rise of big data and cloud computing, deep learning models, especially deep neural networks, arise as a new paradigm for many problems, including image classification, object detection and natural language processing. In recent years, deep learning models are also applied for time series classification and show superiority over traditional models. However, the previous evaluation is usually limited to a small number of datasets and lack of significance analysis. In this study, we give a comprehensive comparison between nearest neighbor and deep learning models. Specifically, we compare 1-NN classifiers with eight different distance measures and three state-of-the-art deep learning models on 128 time series datasets. Our results indicate that deep learning models are not significantly better than 1-NN classifiers with edit distance with real penalty and dynamic time warping.

Unsupervised Learning of Audio Segment Representations using Sequence-to-sequence Recurrent Neural Networks

Abstract: The vector representations of fixed dimensionality for words (in text) offered by Word2Vec have been shown to be very useful in many application scenarios, in particular due to the semantic information they carry. This paper proposes a parallel version, the Audio Word2Vec. It offers the vector representations of fixed dimensionality for variable-length audio segments. These vector representations are shown to describe the sequential phonetic structures of the audio segments to a good degree, with very attractive real world applications such as query-by-example Spoken Term Detection (STD). In this STD application, the proposed approach significantly outperformed the conventional Dynamic Time Warping (DTW) based approaches at significantly lower computation requirements. We propose unsupervised learning of Audio Word2Vec from audio data without human annotation using Sequence-to-sequence Audoencoder (SA). SA consists of two RNNs equipped with Long Short-Term Memory (LSTM) units: the first RNN (encoder) maps the input audio sequence into a vector representation of fixed dimensionality, and the second RNN (decoder) maps the representation back to the input audio sequence. The two RNNs are jointly trained by minimizing the reconstruction error. Denoising Sequence-to-sequence Autoencoder (DSA) is furthered proposed offering more robust learning.