A Neural Network Based Fault Management Scheme for Reliable Image Processing
Open accessThe DOI matches its Crossref record; 32 of the 34 references it cites were found in a registry or an index.
Abstract
Traditional reliability approaches introduce relevant costs to achieve unconditional correctness during data processing. However, many application environments are inherently tolerant to a certain degree of inexactness or inaccuracy. In this article, we focus on the practical scenario of image processing in space, a domain where faults are a threat, while the applications are inherently tolerant to a certain degree of errors. We first introduce the concept of usability of the processed image to relax the traditional requirement of unconditional correctness, and to limit the computational overheads related to reliability. We then introduce our new flexible and lightweight fault management methodology for inaccurate application environments. A key novelty of our scheme is the utilization of neural networks to reduce the costs associated with the occurrence and the detection of faults. Experiments on two aerospace image processing case studies show overall time savings of 14.89 and 34.72 percent for the two applications, respectively, as compared with the baseline classical Duplication with Comparison scheme.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“As discussed in [13], the classical strategy based on a bit-wise comparison of the results against the golden counterpart to classify the experiment as correct versus error is here not effective due to the approximate and inexact nature of CNN applications.”
From Fast and Accurate Error Simulation for CNNs Against Soft Errors (2023) · cites this paper as [13]
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“We train the defined CNNs with the Adam optimizer [25], that exploits the benefits of both AdaGrad [26] and RMSProp [27] strategies.”
Method(2014) · Cited by 33,419
(2010) · Cited by 8,261
ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
Ma, N., Zhang, X., Zheng, H.-T., & Sun, J. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-01264-9_8 · Cited by 7,356 · cited 1×
“We follow the standard and widely adopted guidelines for CNN architecture definition [20] still keeping in mind that there is no “golden” network architecture that is guaranteed to work in the best way for a specific task.”
MethodRecent advances in convolutional neural networks
Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., Liu, T., Wang, X., Wang, G., Cai, J., et al. (2018) · Pattern Recognition · doi:10.1016/j.patcog.2017.10.013 · Cited by 6,576
MnasNet: Platform-Aware Neural Architecture Search for Mobile
Tan, M., Chen, B., Pang, R., Vasudevan, V., Sandler, M., Howard, A., & Le, Q. V. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00293 · Cited by 3,416 · cited 1×
“One can also potentially utilize some of the existing neural architecture search algorithms [34], [35] for finding an appropriate neural network topology for the SC.”
BackgroundThiNet: A Filter Level Pruning Method for Deep Neural Network Compression
Luo, J.-H., Wu, J., & Lin, W. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.541 · Cited by 1,832 · cited 1×
“We applied the following two algorithms [32]:”
A Theoretical Analysis of Feature Pooling in Visual Recognition
Boureau, Y.-L., Ponce, J., & LeCun, Y. (2010) · Cited by 1,660
Overfitting in Neural Nets: Backpropagation, Conjugate Gradient, and Early Stopping
Caruana, R., Lawrence, S., & Giles, C. L. (2000) · neural information processing systems · Cited by 1,318
A Survey of Techniques for Approximate Computing
Mittal, S. (2016) · ACM Computing Surveys · doi:10.1145/2893356 · Cited by 1,058 · cited 1×
“, as in machine learning algorithms), and iii) their output images may be used by a human, whose perceptual limitations provide resiliency to a certain level of inexactness [4].”
BackgroundUnderstanding and Simplifying One-Shot Architecture Search
Bender, G., Kindermans, P.-J., Zoph, B., Vasudevan, V., & Le, Q. V. (2018) · International Conference on Machine Learning · Cited by 834
A Review on Recent Development of Spacecraft Attitude Fault Tolerant Control System
Yin, S., Xiao, B., Ding, S. X., & Zhou, D. (2016) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2016.2530789 · Cited by 430 · cited 1×
“To this aim, hardware/ software Duplication with Comparison (DWC) and Triple Modular Redundancy (TMR) are utilized to achieve fault detection and fault tolerance, respectively [2].”
MethodDeepSat: a learning framework for satellite imagery
Basu, S., Ganguly, S., Mukhopadhyay, S., DiBiano, R., Karki, M., & Nemani, R. (2015) · Proceedings of the 23rd SIGSPATIAL International Conference on Advances in Geographic Information Systems · doi:10.1145/2820783.2820816 · Cited by 303 · cited 1×
“For the second case study we used as input dataset a subset of 20,000 images from the SAT-4 database [33].”
MethodSoft digital signal processing
Hegde, R., & Shanbhag, N. R. (2001) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/92.974895 · Cited by 276 · cited 1×
“Floating Point Unit [13] and for Digital Signal Processing circuits [14], [15].”
BackgroundReliable low-power digital signal processing via reduced precision redundancy
Shim, B., Sridhara, S. R., & Shanbhag, N. R. (2004) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2004.826201 · Cited by 210 · cited 1×
“Other approaches [11], [12] acting at RTL are based on the reduction of the numeric precision in the DWC and TMR schemes.”
MethodQuantifying the Accuracy of High-Level Fault Injection Techniques for Hardware Faults
Wei, J., Thomas, A., Li, G., & Pattabiraman, K. (2014) · 2014 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks · doi:10.1109/dsn.2014.2 · Cited by 182 · cited 4×
“We performed fault injection experiments with LLFI in the four steps of the building identification application.”
BackgroundMethodInfluentialRegularization Techniques for Learning with MatricesPreprint
Kakade, S. M., Shalev‐Shwartz, S., & Tewari, A. (2009) · arXiv (Cornell University) · doi:10.48550/arxiv.0910.0610 · Cited by 162 · cited 1×
“To prevent overfitting, Early Stopping [28], Dropout [24] andRidge Regularization [29] are employed.”
MethodConditional Deep Learning for Energy-Efficient and Enhanced Pattern Recognition
Panda, P., Sengupta, A., & Roy, K. (2016) · Proceedings of the 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE) · doi:10.3850/9783981537079_0819 · Cited by 142
Energy-efficient soft error-tolerant digital signal processing
Shim, B., & Shanbhag, N. R. (2006) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2006.874359 · Cited by 136 · cited 1×
“Floating Point Unit [13] and for Digital Signal Processing circuits [14], [15].”
BackgroundGeneralizing Pooling Functions in Convolutional Neural Networks: Mixed, Gated, and Tree
Lee, C., Gallagher, P. W., & Zhuowen Tu, A. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1509.08985 · Cited by 107 · cited 1×
“of the CNN is a fully-connected layer that performs the final classification [21], [22], [23].”
Approximate logic circuits for low overhead, non-intrusive concurrent error detection
Choudhury, M. R., & Mohanram, K. (2008) · Design, Automation, and Test in Europe · doi:10.1145/1403375.1403593 · Cited by 80 · cited 1×
“ment schemes have been considered in [9], starting from the DWC, and other are presented in [10], devoted to reduce the size of the checker/voter circuits.”
BackgroundDesigning adaptive neural networks for energy-constrained image classification
Stamoulis, D., Chin, T.-W. (., Prakash, A. K., Fang, H., Sajja, S., Bognar, M., & Marculescu, D. (2018) · Proceedings of the International Conference on Computer-Aided Design · doi:10.1145/3240765.3240796 · Cited by 63 · cited 1×
“A similar design choice is considered in the context of multi-stage conditional neural networks [30], [31].”
BackgroundExploring the use of approximate TMR to mask transient faults in logic with low area overhead
Gomes, I. A. C., Martins, M. G. A., Reis, A. I., & Kastensmidt, F. L. (2015) · Microelectronics Reliability · doi:10.1016/j.microrel.2015.06.125 · Cited by 59 · cited 1×
“Various approximate/inexact TMR schemes have been proposed at the logic level [7], [8], [9], [10].”
BackgroundSanchez-Clemente, A., Entrena, L., Hrbáček, R., & Sekanina, L. (2016) · IEEE Transactions on Reliability · doi:10.1109/tr.2016.2604918 · Cited by 51 · cited 1×
“Various approximate/inexact TMR schemes have been proposed at the logic level [6], [7], [8].”
BackgroundError Resilient GPU Accelerated Image Processing for Space Applications
Davidson, R. L., & Bridges, C. (2018) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2018.2812853 · Cited by 43 · cited 1×
“For example, in the aerospace domain, the navigation system of a satellite is a mission-critical application while the payload processing applications are not [1].”
BackgroundMajority Voting-Based Reduced Precision Redundancy Adders
Ullah, A., Reviriego, P., Pontarelli, S., & Maestro, J. A. (2017) · IEEE Transactions on Device and Materials Reliability · doi:10.1109/tdmr.2017.2781186 · Cited by 32 · cited 1×
“Other approaches [11], [12] acting at RTL are based on the reduction of the numeric precision in the DWC and TMR schemes.”
MethodImproving approximate-TMR using multi-objective optimization genetic algorithm
Albandes, I., Serrano-Cases, A., Sanchez-Clemente, A., Martins, M. G. A., Martínez-Álvarez, A., Cuenca-Asensi, S., & Kastensmidt, F. L. (2018) · IEEE Latin American Test Symposium (LATS) · doi:10.1109/latw.2018.8349665 · Cited by 23 · cited 1×
“Various approximate/inexact TMR schemes have been proposed at the logic level [6], [7], [8].”
BackgroundVariable-Length Limited Feedback Beamforming in Multiple-Antenna Fading Channels
Koyuncu, E., & Jafarkhani, H. (2014) · IEEE Transactions on Information Theory · doi:10.1109/tit.2014.2357811 · Cited by 21 · cited 1×
“It has also been observed in different contexts that designing a system to provide an acceptable/usable output rather than the best-possible/correct output can yield significant gains in terms of the system’s implementation costs and complexity [18], [19].”
BackgroundSelf-citationTwo Approximate Voting Schemes for Reliable Computing
Chen, K., Han, J., & Lombardi, F. (2017) · IEEE Transactions on Computers · doi:10.1109/tc.2017.2653780 · Cited by 21 · cited 1×
“ment schemes have been considered in [9], starting from the DWC, and other are presented in [10], devoted to reduce the size of the checker/voter circuits.”
BackgroundReduced Precision Checking for a Floating Point Adder
Eibl, P. J., Cook, A. D., & Sorin, D. J. (2009) · Proceedings · doi:10.1109/dft.2009.22 · Cited by 14 · cited 2×
“Floating Point Unit [13] and for Digital Signal Processing circuits [14], [15].”
BackgroundKoyuncu, E., Zou, X., & Jafarkhani, H. (2018) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2018.2864204 · Cited by 10
Self-citationA Smart Fault Detection Scheme for Reliable Image Processing Applications
Biasielli, M., Bolchini, C., Cassano, L., & Miele, A. (2019) · Design, Automation, and Test in Europe · doi:10.23919/date.2019.8714945 · Cited by 4 · cited 9×
“This architecture includes both the traditionalDWCscheme (corresponding to setting a 1⁄4 0% and having no CNN block), and the solution presented in [5] (corresponding to setting a 1⁄4 0% and having a CNNblock, i.”
BackgroundMethodResultInfluentialSelf-citation
Research works printed but not linked: 1
- abadi, 2016, TensorFlow: Large-scale machine learning on heterogeneous systems, Proc USENIX Symp on Operating System Design and Implementation, 265
Cited by
Furano, G., Meoni, G., Dunne, A., Moloney, D., Ferlet-Cavrois, V., Tavoularis, A., Byrne, J., Buckley, L., Psarakis, M., Voss, K.-O., et al. (2020) · IEEE Aerospace and Electronic Systems Magazine · doi:10.1109/maes.2020.3008468 · Cited by 259
“On the contrary, for those applications requiring reprocessing the image, the detection of “slightly corrupted” images allows saving time due to image processing pipeline [45].”
BackgroundMethodRech, P. (2024) · IEEE Transactions on Nuclear Science · doi:10.1109/tns.2024.3349956 · Cited by 75
“A similar approach has been later applied to image processing pipelines [176], [177] with similar performances.”
MethodApproximation-based Fault Tolerance in Image Processing Applications
Biasielli, M., Bolchini, C., Cassano, L., Mazzeo, A., & Miele, A. (2021) · IEEE Transactions on Emerging Topics in Computing · doi:10.1109/tetc.2021.3100623 · Cited by 21
Self-citationQian, Y., & Qiao, Y. (2025) · Communications in Nonlinear Science and Numerical Simulation · doi:10.1016/j.cnsns.2025.108687 · Cited by 16
Wu, X. H., Liu, S., Wang, Y., & Liu, Z. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.07.042 · Cited by 15
Metadata
- Authors
- Biasielli, M., Bolchini, C., Cassano, L., Koyuncu, E., & Miele, A.
- Affiliations
- Politecnico di Milano (IT)
- University of Illinois Chicago (US)
- Venue
- IEEE Transactions on Computers
- Published
- 2020-05-01
- Volume
- 69
- Issue
- 5
- Pages
- 764-776
- ISSN
- 0018-9340 (ISSN-L)
- 1557-9956
- 2326-3814
- OpenAlex
- W2999211665
- Cited by
- 28
- References
- 34
- Citations by year
- 2026: 5 · 2025: 1 · 2024: 5 · 2023: 2 · 2022: 2 · 2021: 8 · 2020: 5
- FWCI
- 2.00
- Type
- journal article
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- Locations
- IEEE Transactions on Computers (publishedVersion)
- Virtual Community of Pathological Anatomy (University of Castilla La Mancha) (submittedVersion, other-oa)
- Funders
- University of Illinois at Chicago
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed systems and fault tolerance
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Radiation Effects in Electronics
- Physical Sciences › Engineering › Control and Systems Engineering › Fault Detection and Control Systems
- Keywords
- fault management, artificial neural networks, fault tolerance, fault detection, fault recovery
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.