M.L. Hilton

Bio: M.L. Hilton is an academic researcher from University of South Carolina. The author has contributed to research in topics: Computer science & Deep learning. The author has an hindex of 1, co-authored 1 publications receiving 434 citations.

Wavelet and wavelet packet compression of electrocardiograms

Abstract: Wavelets and wavelet packets have recently emerged as powerful tools for signal compression. Wavelet and wavelet packet-based compression algorithms based on embedded zerotree wavelet (EZW) coding are developed for electrocardiogram (ECG) signals, and eight different wavelets are evaluated for their ability to compress Holter ECG data. Pilot data from a blind evaluation of compressed ECG's by cardiologists suggest that the clinically useful information present in original ECG signals is preserved by 8:1 compression, and in most cases 16:1 compressed ECG's are clinically useful.

Utility of machine learning for segmenting camera trap time‐lapse recordings

Abstract: Camera trap time-lapse recordings can collect vast amounts of data on wildlife in their natural settings. Transforming these data into information useful to ecologists is a major challenge. Machine learning techniques show promise for becoming important tools in the cost-effective analysis of camera trap data, but only if they become readily available to researchers without requiring advanced computing skills and resources. We present a new suite of software tools that reduce the amount of human effort needed to segment time-lapse, camera trap recordings in preparation for analysis. The tools incorporate a convolutional neural network trained to detect a focal species and to generate a draft video segmentation indicating the ranges of time when the focal species is present. We evaluated the utility of our neural network by comparing manual and automatic segmentations of 64 time-lapse recordings of gopher tortoise (Gopherus polyphemus) burrows, recorded in Pinellas County, Florida, USA between 25 November 2020 and 30 November 2020. The neural network correctly found 130 of the 145 segments containing tortoises (89.7%), whereas student graders found 135 segments (93.1%). A year of experience using the new software suite in an ongoing study of gopher tortoises deploying 12 camera traps indicates one person, assisted by machine learning algorithms, can segment a week's worth of time-lapse recordings—11.5 hours of standard-speed video—in under 3 hours. We concluded that the use of machine learning algorithms is practical and allows researchers to process large volumes of time-lapse data with minimal human effort.

An Image Processing Pipeline for Camera Trap Time-Lapse Recordings

Abstract: A new open-source image processing pipeline for analyzing camera trap time-lapse recordings is described. This pipeline includes machine learning models to assist human-in-the-loop video segmentation and animal reidentification. We present some performance results and observations on the utility of this pipeline after using it in a year-long project studying the spatial ecology and social behavior of the gopher tortoise.

Wavelets on Graphs via Spectral Graph Theory

Abstract: We propose a novel method for constructing wavelet transforms of functions defined on the vertices of an arbitrary finite weighted graph. Our approach is based on defining scaling using the the graph analogue of the Fourier domain, namely the spectral decomposition of the discrete graph Laplacian $\L$. Given a wavelet generating kernel $g$ and a scale parameter $t$, we define the scaled wavelet operator $T_g^t = g(t\L)$. The spectral graph wavelets are then formed by localizing this operator by applying it to an indicator function. Subject to an admissibility condition on $g$, this procedure defines an invertible transform. We explore the localization properties of the wavelets in the limit of fine scales. Additionally, we present a fast Chebyshev polynomial approximation algorithm for computing the transform that avoids the need for diagonalizing $\L$. We highlight potential applications of the transform through examples of wavelets on graphs corresponding to a variety of different problem domains.

Compressed Sensing for Real-Time Energy-Efficient ECG Compression on Wireless Body Sensor Nodes

Abstract: Wireless body sensor networks (WBSN) hold the promise to be a key enabling information and communications technology for next-generation patient-centric telecardiology or mobile cardiology solutions. Through enabling continuous remote cardiac monitoring, they have the potential to achieve improved personalization and quality of care, increased ability of prevention and early diagnosis, and enhanced patient autonomy, mobility, and safety. However, state-of-the-art WBSN-enabled ECG monitors still fall short of the required functionality, miniaturization, and energy efficiency. Among others, energy efficiency can be improved through embedded ECG compression, in order to reduce airtime over energy-hungry wireless links. In this paper, we quantify the potential of the emerging compressed sensing (CS) signal acquisition/compression paradigm for low-complexity energy-efficient ECG compression on the state-of-the-art Shimmer WBSN mote. Interestingly, our results show that CS represents a competitive alternative to state-of-the-art digital wavelet transform (DWT)-based ECG compression solutions in the context of WBSN-based ECG monitoring systems. More specifically, while expectedly exhibiting inferior compression performance than its DWT-based counterpart for a given reconstructed signal quality, its substantially lower complexity and CPU execution time enables it to ultimately outperform DWT-based ECG compression in terms of overall energy efficiency. CS-based ECG compression is accordingly shown to achieve a 37.1% extension in node lifetime relative to its DWT-based counterpart for “good” reconstruction quality.