The degree of sensitisation is a reliable measure of the susceptibility to intergranular corrosion for Al-Mg and Al-Mg-Mn marine-grade alloys (Al 5XXX series). Measuring the degree of sensitisation is an exhaustive and expensive operation, as it requires destructive sectioning and outsourcing the measurement to a laboratory to determine the necessary corrosion risk level. This paper proposes an alternative method to predict degree of sensitisation using a convolutional neural network trained on microstructural images of AA5456 alloy specimens etched with phosphoric acid to reveal β-phase precipitates along grain boundaries. Two convolutional neural network architectures are proposed to predict the degree of sensitisation of AA5456 specimens from 490 microstructural images at 200× and 500× magnification. An overall accuracy of 83% for a continuous prediction convolutional neural network model and 87% for a classification prediction convolutional neural network model. The high classification accuracy reflects the success of the augmentation process, which increased the dataset to 9000 images with an equal distribution of degree of sensitisation targets.
SuttonF. Investigation of corrosion and stress-corrosion properties of nonheat-treatable aluminum alloys in marine environments. Technical Report 910037K, United States Naval Engineering Experiment Station, 1959.
2.
LamirandeE. Aluminum alloys – marine environment – corrosion. Technical Memorandum EA152-290-71, Naval Underwater Systems Center, 1971.
3.
VassilarosM. Five year corrosion study of aluminum magnesium alloys. Technical Report DTNSRDC/SME-84-67, David W. Taylor Naval Ship Research and Development Center, 1984.
4.
BayRSchrockDAkmanA, et al.The effect of sensitization and fatigue loading frequency on corrosion fatigue of AA5083-H131. Int J Fatigue2019; 124: –9.
5.
BaerDRWindischCFEngelhardMH, et al.Influence of mg on the corrosion of al. J Vac Sci Technol1999; 18: 131–136.
6.
LinYKWangSHChenRY, et al.The effect of heat treatment on the sensitized corrosion of the 5386-H116 al-mg alloy. Materials (Basel)2017; 10: 75.
7.
ASTM International. Standard test method for determining the susceptibility to intergranular corrosion of 5xxx series aluminum alloys by mass loss after exposure to nitric acid (NAMLT test). Standard G67-18, ASTM International, West Conshohocken, PA, 2018.
8.
KantVKumarBV. Deep learning-based classification of lung and colon cancer subtypes using histology images. In: 2025 4th OPJU international technology conference (OTCON) on smart computing for innovation and advancement in industry 5.0, pp.1–5, 2025. doi: 10.1109/OTCON65728.2025.11070738.
9.
JiSXuWYangM, et al.3D Convolutional neural networks for human action recognition. IEEE Trans Pattern Anal Mach Intell2013; 35: 221–231.
10.
TiviveFBouzerdoumA. A new class of convolutional neural networks (Siconnets) and their application of face detection. In: Proceedings of the international joint conference on neural networks (IJCNN), pp.2571–2575, 2003. doi: 10.1109/IJCNN.2003.1223742.
11.
FuYDowneyAYuanL, et al.Machine learning algorithms for defect detection in metal laser-based additive manufacturing: a review. J Manuf Process2022; 75: 693–710.
12.
O'SheaKNashR. An introduction to convolutional neural networks. arXiv preprint arXiv:1511.08458, 2015. doi: 10.48550/ARXIV.1511.08458.
13.
CireşanDMeierUSchmidhuberJ. Multi-column deep neural networks for image classification. arXiv preprint arXiv:1202.2745, 2012. doi: 10.48550/arXiv.1202.2745.
14.
KrizhevskyASutskeverIHintonG. Imagenet classification with deep convolutional neural networks. Commun ACM2017; 60: 84–90.
15.
CoelhoLZhangDVan IngelgemYV, et al.Reviewing machine learning of corrosion prediction in a data-oriented perspective. npj Mater Degrad2022; 6: 8.
16.
CuiYJiaMLinTY, et al.Class-balanced loss based on effective number of samples. arXiv preprint arXiv:1901.05555, 2019. doi: 10.48550/arXiv.1901.05555.
17.
KingmaDBaJ. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2017. doi: 10.48550/arXiv.1412.6980.
18.
RameshGSrihariWBharadwajA, et al.Kannada ImageNet: A dataset for image classification in Kannada. In: 2021 International conference on computer communication and informatics (ICCCI), pp.1–4, 2021. doi: 10.1109/ICCCI50826.2021.9402356.
19.
SunCShrivastavaASinghS, et al.Revisiting unreasonable effectiveness of data in the deep learning era. arXiv preprint arXiv:1707.02968, 2017. doi: 10.48550/arXiv.1707.02968.
20.
RolnickDVeitABelongieS, et al.Deep learning is robust to massive label noise, arXiv preprint arXiv:1705.10694, 2018. doi: 10.48550/arXiv.1705.10694.
21.
LeCunYBottouLBengioY, et al.Gradient-based learning applied to document recognition. Proc IEEE1998; 86: 2278–2324.
22.
ZhangWZhaiGYangX, et al.Hierarchical features fusion for image aesthetics assessment. In: Proceedings of the 2019 IEEE international conference on image processing (ICIP), pp.3771–3775, 2019. doi: 10.1109/ICIP.2019.8803599.
23.
BaskettWShyuCR. Addressing small sample size and improving interpretability in biomedical applications with adaptive complexity deep neural networks. In: Proceedings of the 2025 IEEE 13th international conference on healthcare informatics (ICHI), pp.30–38, 2025. doi: 10.1109/ICHI64645.2025.00012.
24.
ZhangXWangPLiJ. Contrast enhancement of metallographic microstructure images using adaptive histogram equalization techniques. Graph Models2019; 105: 101206.
25.
ChowdhurySRahmanFSarkerM. Improving CNN robustness under noisy imaging conditions. Mach Vis Appl2020; 31: 05.
26.
RahmanMChakrabortySFukudaY. CNN Ensemble techniques for robust image recognition. SN Comput Sci2021; 2: 63.
27.
KumarRSinghA. Hybrid ensemble learning for image-based corrosion analysis. IMA J Math Control Inf Early Access2024. doi: 10.1002/ima.22761
28.
LiuHDasSKimT. Deep ensemble CNNs for materials classification. Circuits Syst Signal Process2023; 42: 2564.
