Preprint
Computer Vision

1D convolutional neural networks and applications: A survey

Kiranyaz, Mustafa Serkan(Qatar University), Onur Avcı(Iowa State University), Osama Abdeljaber(Linnaeus University), Türker İnce(İzmir University of Economics), Moncef Gabbouj(Tampere University), Daniel J. Inman(University of Michigan)
January 1, 2021Qatar University QSpace (Qatar University)2,733 citations
arXivDOI

2.7k

Citations

66

Influential Citations

Qatar University QSpace (Qatar University)

Venue

2021

Year

Abstract

During the last decade, Convolutional Neural Networks (CNNs) have become the de facto standard for various Computer Vision and Machine Learning operations. CNNs are feed-forward Artificial Neural Networks (ANNs) with alternating convolutional and subsampling layers. Deep 2D CNNs with many hidden layers and millions of parameters have the ability to learn complex objects and patterns providing that they can be trained on a massive size visual database with ground-truth labels. With a proper training, this unique ability makes them the primary tool for various engineering applications for 2D signals such as images and video frames. Yet, this may not be a viable option in numerous applications over 1D signals especially when the training data is scarce or application specific. To address this issue, 1D CNNs have recently been proposed and immediately achieved the state-of-the-art performance levels in several applications such as personalized biomedical data classification and early diagnosis, structural health monitoring, anomaly detection and identification in power electronics and electrical motor fault detection. Another major advantage is that a real-time and low-cost hardware implementation is feasible due to the simple and compact configuration of 1D CNNs that perform only 1D convolutions (scalar multiplications and additions). This paper presents a comprehensive review of the general architecture and principals of 1D CNNs along with their major engineering applications, especially focused on the recent progress in this field. Their state-of-the-art performance is highlighted concluding with their unique properties. The benchmark datasets and the principal 1D CNN software used in those applications are also publicly shared in a dedicated website. While there has not been a paper on the review of 1D CNNs and its applications in the literature, this paper fulfills this gap.

Analysis

Why This Paper Matters

This paper is significant because it is the first comprehensive review of 1D Convolutional Neural Networks (CNNs), a class of models that has been largely overshadowed by their 2D counterparts in computer vision. While 2D CNNs dominate image and video processing, many real-world signals—such as ECG, vibration, and power signals—are inherently 1D. The authors argue that 1D CNNs offer a more natural and efficient approach for these applications, especially when training data is scarce or application-specific. By consolidating the state of the art, this survey provides a crucial reference that validates the effectiveness of 1D CNNs and encourages their adoption in fields like biomedical engineering and structural health monitoring.

The paper also highlights a practical advantage: 1D CNNs are computationally simpler and more compact, enabling real-time and low-cost hardware implementations. This is particularly important for edge devices and embedded systems, where resource constraints are critical. The survey's publication in 2021, with over 2,700 citations, underscores its influence and the growing interest in efficient neural architectures for signal processing.

Technical Contributions

The paper's main technical contributions include:

  • Comprehensive Architecture Review: It details the general architecture of 1D CNNs, including convolutional and subsampling layers, and explains how they differ from 2D CNNs.
  • Application Taxonomy: It categorizes major engineering applications, such as biomedical data classification, structural health monitoring, and electrical motor fault detection, providing a structured overview.
  • Performance Benchmarking: It highlights state-of-the-art results achieved by 1D CNNs in these domains, often outperforming traditional methods and even 2D CNNs when adapted to 1D signals.
  • Resource Sharing: The authors publicly share benchmark datasets and principal 1D CNN software on a dedicated website, facilitating reproducibility and further research.

Results

The paper does not present new experimental results but synthesizes findings from numerous studies. It reports that 1D CNNs achieve state-of-the-art performance in personalized biomedical data classification, early diagnosis, structural health monitoring, anomaly detection in power electronics, and electrical motor fault detection. A key result is that 1D CNNs can achieve high accuracy even with limited training data, which is a common challenge in these applications. Additionally, the compact configuration of 1D CNNs enables real-time inference on low-cost hardware, a significant advantage over deeper 2D models.

Significance

This survey has had a broad impact on the AI community by legitimizing 1D CNNs as a powerful tool for signal processing. It has inspired further research into efficient architectures and has practical implications for deploying AI in resource-constrained environments. By providing a centralized resource, the paper lowers the barrier to entry for practitioners and encourages cross-disciplinary adoption. As the field moves toward edge AI and personalized medicine, 1D CNNs are likely to play an increasingly important role, and this paper serves as a foundational reference for that progress.

1d-cnnsurveybiomedicalstructural-health-monitoringanomaly-detectionfault-detectionreal-timecompact-models
BenchmarksComputer VisionNeural Networks

Related Papers