Search Advanced

Mathematical Biosciences and Engineering

2024, Volume 21, Issue 6: 6190-6224. doi: 10.3934/mbe.2024272
Previous Article Next Article
Review Special Issues

Image data augmentation techniques based on deep learning: A survey

  • Engineering Training Center, Putian University, Putian 351100, China
  • Received: 09 February 2024 Revised: 19 May 2024 Accepted: 31 May 2024 Published: 12 June 2024

  • In recent years, deep learning (DL) techniques have achieved remarkable success in various fields of computer vision. This progress was attributed to the vast amounts of data utilized to train these models, as they facilitated the learning of more intricate and detailed feature information about target objects, leading to improved model performance. However, in most real-world tasks, it was challenging to gather sufficient data for model training. Insufficient datasets often resulted in models prone to overfitting. To address this issue and enhance model performance, generalization ability, and mitigate overfitting in data-limited scenarios, image data augmentation methods have been proposed. These methods generated synthetic samples to augment the original dataset, emerging as a preferred strategy to boost model performance when data was scarce. This review first introduced commonly used and highly effective image data augmentation techniques, along with a detailed analysis of their advantages and disadvantages. Second, this review presented several datasets frequently employed for evaluating the performance of image data augmentation methods and examined how advanced augmentation techniques can enhance model performance. Third, this review discussed the applications and performance of data augmentation techniques in various computer vision domains. Finally, this review provided an outlook on potential future research directions for image data augmentation methods.

    Citation: Wu Zeng. Image data augmentation techniques based on deep learning: A survey[J]. Mathematical Biosciences and Engineering, 2024, 21(6): 6190-6224. doi: 10.3934/mbe.2024272

    Related Papers:

    [1] Marina A. Darenskaya, Liubov I. Kolesnikova, Liubov V. Rychkova, Olga V. Kravtsova, Natalya V. Semenova, Sergei I. Kolesnikov . Relationship between lipid metabolism state, lipid peroxidation and antioxidant defense system in girls with constitutional obesity. AIMS Molecular Science, 2021, 8(2): 117-126. doi: 10.3934/molsci.2021009
    [2] Naba Hasan, Waleem Ahmad, Feroz Alam, Mahboob Hasan . Ferroptosis-molecular mechanisms and newer insights into some diseases. AIMS Molecular Science, 2023, 10(1): 22-36. doi: 10.3934/molsci.2023003
    [3] Vahid Pouresmaeil, Marwa Mawlood Salman Al-zand, Aida Pouresmaeil, Seyedeh Samira Saghravanian, Masoud Homayouni Tabrizi . Loading diltiazem onto surface-modified nanostructured lipid carriers to evaluate its apoptotic, cytotoxic, and inflammatory effects on human breast cancer cells. AIMS Molecular Science, 2024, 11(3): 231-250. doi: 10.3934/molsci.2024014
    [4] Geovanni Alberto Ruiz-Romero, Carolina Álvarez-Delgado . Effects of estrogens in mitochondria: An approach to type 2 diabetes. AIMS Molecular Science, 2024, 11(1): 72-98. doi: 10.3934/molsci.2024006
    [5] Fatima Al-Hashimi, Salvador J. Diaz-Cano . Multi-target analysis of neoplasms for the evaluation of tumor progression: stochastic approach of biologic processes. AIMS Molecular Science, 2018, 5(1): 14-62. doi: 10.3934/molsci.2018.1.14
    [6] Jacob Fosso-Tande, Cody Black, Stephen G. Aller, Lanyuan Lu, Ronald D. Hills Jr . Simulation of lipid-protein interactions with the CgProt force field. AIMS Molecular Science, 2017, 4(3): 352-369. doi: 10.3934/molsci.2017.3.352
    [7] Vittorio Emanuele Bianchi, Giancarlo Falcioni . Reactive oxygen species, health and longevity. AIMS Molecular Science, 2016, 3(4): 479-504. doi: 10.3934/molsci.2016.4.479
    [8] Nguyen Hoai Nguyen . HY5, an integrator of light and temperature signals in the regulation of anthocyanins biosynthesis in Arabidopsis. AIMS Molecular Science, 2020, 7(2): 70-81. doi: 10.3934/molsci.2020005
    [9] Teresa M. Fonovich . Phospholipid synthetic and turnover pathways elicited upon exposure to different xenobiotics. AIMS Molecular Science, 2020, 7(3): 211-228. doi: 10.3934/molsci.2020010
    [10] Siddig Ibrahim Abdelwahab, Manal Mohamed Elhassan Taha, Adel S. Al-Zubairi, Ahmad Syahida, Lee KaHeng, Putri Narrima, Rozana Othman, Hassan Ahmad Alfaifi, Amal Hamdan Alzahrani . Anti-inflammatory and antioxidant properties of bark and fruit extracts of Faidherbia albida (Delile) A. Chev: A perspective from bio-prospecting assays to scientometric approach. AIMS Molecular Science, 2024, 11(3): 262-276. doi: 10.3934/molsci.2024016

  • In recent years, deep learning (DL) techniques have achieved remarkable success in various fields of computer vision. This progress was attributed to the vast amounts of data utilized to train these models, as they facilitated the learning of more intricate and detailed feature information about target objects, leading to improved model performance. However, in most real-world tasks, it was challenging to gather sufficient data for model training. Insufficient datasets often resulted in models prone to overfitting. To address this issue and enhance model performance, generalization ability, and mitigate overfitting in data-limited scenarios, image data augmentation methods have been proposed. These methods generated synthetic samples to augment the original dataset, emerging as a preferred strategy to boost model performance when data was scarce. This review first introduced commonly used and highly effective image data augmentation techniques, along with a detailed analysis of their advantages and disadvantages. Second, this review presented several datasets frequently employed for evaluating the performance of image data augmentation methods and examined how advanced augmentation techniques can enhance model performance. Third, this review discussed the applications and performance of data augmentation techniques in various computer vision domains. Finally, this review provided an outlook on potential future research directions for image data augmentation methods.


    1.   Introduction

    The relevance of studying diseases of the hepatopancreatoduodenal region is currently quite high due to the widespread prevalence and increasing number of patients [1]. Mechanical jaundice (MJ) is the most common among these diseases and is found in 12–45% of cases, mostly men [2]. MJ, or bile duct blockage code K83.1 (ICD 10), is a complex of clinical and morphological manifestations that develop when the bile pathway is disturbed and bile flow from the liver to the duodenum through the main bile ducts is stopped [3]. MJ may be of benign or malignant etiology [1]. The causes of benign processes are diverse and include cholelithiasis, pancreatitis, stenosis of the large duodenal nipple, cicatricial narrowing of the extrahepatic bile ducts, congenital anomalies of the main bile ducts, and parasitic liver diseases [4]. Malignant neoplasms of the pancreas, large duodenal papilla, bile ducts, and liver cancer (primary and metastatic) account for about 20% of MJ causes [5]. Complications of MJ include biliary tract lesions and liver dysfunction, along with systemic complications, including multiple organ failure syndrome [6].

    The main pathogenetic mechanisms of development of complications in MJ are: cholestasis with subsequent development of hepatic-cellular insufficiency; disruption of detoxifying mechanisms; immune system imbalance, and oxidative stress reactions [3],[7],[8]. Despite advances in preoperative assessment and postoperative care, intervention, especially surgical intervention, and MJ management remain associated with high rates of morbidity and mortality, mainly due to sepsis and liver dysfunction [9]. The main trigger mechanism for hepatocyte damage and development of liver failure is hypoxia of liver cells due to increased intra-flow pressure, as well as impaired perfusion of liver tissues [10]. Development of circulatory hypoxia can lead to an intensification of oxidative stress reactions and a lack of antioxidant factors [8]. It has been shown experimentally that violation of the intestinal barrier in MJ is associated with a high frequency of oxidative stress reactions in the intestine [11],[12]. Altered expression of intestinal tight junctions and increased cell apoptosis are accompanied by significant changes in the oxidative state of the intestine, an additional important factor contributing to intestinal wall damage in MJ [13],[14]. Due to active involvement of oxidative stress reactions in MJ genesis, it is relevant to study the parameters of this process in patients with MJ before and after surgical intervention to develop possible recommendations regarding such intervention.

    In this regard, this study assesses lipid peroxidation-antioxidant defense components in men with MJ of non-tumor and tumor origin in the context of surgical treatment.

    2.   Materials and methods

    2.1.   Study design

    This study is designed as a prospective cross-sectional observational study. We use data from 47 men (average age 52.02 ± 5.18 years) with mechanical jaundice of non-tumor origin (MJNT) and 45 men with mechanical jaundice of tumor origin (MJT) (stages I-II of the tumor process) (average age 53.02 ± 4.8 years). The control group consisted of healthy men (n = 100, average age 48.7 ± 3.9 years). The studied clinical groups were comparable in age, body mass index, alcohol consumption, tobacco smoking, physical activity, and drug use (p > 0.05). Criteria for inclusion in an MJ group were: confirmed diagnosis of MJ of non-tumor origin due to gallstone disease or syndrome or MJ of tumor origin, the cause of which was cholangiocarcinoma (stage I–II of neoplasia); total bilirubin levels higher than 3 mg/dL or conjugated bilirubin higher than 2 mg/dL; and ultrasound evidence of extrahepatic bile duct dilatation higher than 12 mm and intrahepatic bile duct dilation higher than 4 mm. Exclusion criteria were: acute cholangitis, parenchymal liver disease, gastrointestinal hemorrhage, prior or concomitant intravenous fluid therapy, heart failure or chronic renal failure, and treatment with diuretics and antihypertensives. Informed consent was obtained from all participants. All patients underwent transcutaneous decompression of the bile ducts to stop jaundice. This study was approved by the ethics committee of the Research Institute of Medical Problems of the North, Krasnoyarsk Science Center Siberian Branch of the Russian Academy of Sciences (Protocol No. 7 of 16.11.2012). When examining patients, ethical principles were observed (according to the Helsinki Declaration of the World Medical Association, 1964, ed. 2013).

    2.2.   Biochemical measurements

    Patient and control group peripheral blood biochemical and oxidative stress parameters were measured on admission to hospital and 7 days after surgical intervention.

    Blood plasma was obtained after blood centrifugation at 3000 g for 5 min at 4°C. Samples were stored at 80°C temperature until laboratory analysis. Dynamic monitoring of patients was carried out throughout the period of hospital admission.

    To assess lipid peroxidation-antioxidant defense content (LPO-AOD) system state, the following component content was determined: conjugated dienes (CDs) [15], thiobarbituric acid reactants (TBARs) [16], levels of glutathione-S-transferase (GST) and glutathione peroxidase (GPO) [17], superoxide dismutase (SOD) [18], catalase [17], and ceruloplasmin (CP) [19]. Values were measured using a Shimadzu RF-1501 spectrofluorophotometer (Shimadzu, Japan).

    2.3.   Statistical procedure

    The Statistica for Windows 8.0 application package (Stat Soft Inc., USA, 2008) was used for statistical processing of research results. Determination of proximity of a sample to the normal distribution was performed using the visual-graphical method, the Kolmogorov-Smirnov agreement criteria, and the Lilliefors and Shapiro-Wilk tests. Analysis of statistical significance of the revealed differences in quantitative data was carried out via Mann–Whitney rank criterion. The critical significance level for testing statistical hypotheses was assumed to be p < 0.05.

    3.   Results

    LPO-AOD system analysis indicated that there was a statistically significant increase in the content of primary LPO products (CDs 1.83 times higher; p = 0.001) and final LPO products (TBARs 1.83 times higher; p < 0.001) in the group of patients with MJNT relative to the control group (Figure 1). At the same time, in this group, there was a significant decrease in the levels of almost all AOD system studied components: SOD (1.25 times lower; p = 0.04), GST (1.25 times lower; p = 0.03), GPO (1.15 times lower; p < 0.001) and CP (1.4 times lower; p = 0.01) relative to the control group. A similar trend was observed in patients with MJT, but in this group, there were more intense changes in parameters (Figure 2). There was an increase in CDs (2.87 times higher; p < 0.001) and TBARs (1.88 times higher; p < 0.001), a decrease in SOD (1.38 times lower; p < 0.001), GST (2.22 times lower; p < 0.001), GPO (1.2 times lower; p < 0.001) and CP (1.79 times lower; p < 0.001) compared to reference values. Differences between groups were observed only in relation to one indicator, CP, the values of which were lower in the group with MJT (1.4 times lower; p < 0.001) (Figure 2).

    Figure 1.  Changes in parameters of LPO-AOD system in men with MJNT before and after surgical intervention; * indicates statistically significant differences compared to control group; indicates statistically significant differences in men with MJNT after surgical intervention compared to group before surgical intervention.
    DownLoad: Full-Size Img PowerPoint

    After surgical intervention, average CDs values in patients with MJNT compared to control group values remained elevated (1.57 times higher; p < 0.001), while values of some components of the AOD system decreased slightly: SOD (1.1 times lower; p = 0.03), GST (1.15 times lower; p = 0.04) and CP (1.29 times lower; p = 0.043) (Figure 1). In the MJNT group post-surgery, relative to pre-surgery data, there was a decrease in TBARs values (1.33 times lower; p = 0.003), an increase in catalase values (1.25 times higher; p = 0.028), and an increase in CP values (1.09 times higher; p = 0.03) (Figure 1).

    A similar trend was observed in the MJT group: an increase in the content of CDs (2.35 times higher; p < 0.001) and TBARs (1.75 times higher; p < 0.001), a decrease in SOD (1.21 times lower; p < 0.001), GST (2 times lower; p = 0.002), GPO (1.2 times lower; p = 0.001) and CP (1.4 times lower; p = 0.003) relative to the control group (Figure 2). In the MJT group post-surgery, relative to pre-surgery data, there was a decrease in CDs values (1.22 times lower; p = 0.037), an increase in catalase values (1.67 times higher; p = 0.0009), and an increase in CP values (1.27 times higher; p = 0.03) (Figure 2).

    Figure 2.  Changes in parameters of LPO-AOD system in men with MJT before and after surgical intervention; * indicates statistically significant differences compared to control group; indicates statistically significant differences in men with MJT after surgical intervention compared to group before surgical intervention.
    DownLoad: Full-Size Img PowerPoint

    4.   Discussion

    This study found that MJNT leads to significant activation of lipid peroxidation processes: increases in both primary (CDs) and final (TBARs) LPO products, at levels more pronounced than in the MJT group. LPO products are classified as toxic compounds that have a damaging effect on various components of the cell, so their increase can have an extremely negative impact on the state of the body as a whole [20]–[22]. Development of oxidative stress in MJ can be initiated by the absence of normal bile flow, which deprives the intestine of its bacteriostatic, neutralizing, and trophic properties [23]. There is an increase in bacterial load and mucosal atrophy, which contributes to translocation of bacteria and endotoxins into the portal circulation [24]. Cholestasis reduces Kupffer-cell clearance ability, which provokes ingress of bacteria and endotoxins further into the systemic bloodstream. A systemic inflammatory reaction develops, which exacerbates intestinal barrier dysfunction and cholestatic liver damage [25],[26]. In complex endotoxemia, high levels of cytokines and bile acids are important stimulators of oxidative damage in various organs, including the intestines. Studies on experimental animals have shown a significant increase in LPO products in the intestinal mucosa, as well as a general imbalance between protein or non-protein thiols [10],[14]. Data on increased intestinal oxidative stress in MJ have also been confirmed in clinical settings [3]. It is known that depending on the duration of the pathological process, oxidative stress can cause both cell death and activation of adaptive defense mechanisms that lead to the new ratio of reactive oxygen species/antioxidants [27].

    This study found a significant decrease in almost all the studied components of the AOD system, SOD, GST, GPO, and CP, in the MJNT group. A similar situation, but with more pronounced changes, was observed in the MJT group. It is believed that even small fluctuations in SOD levels indicate a shift in metabolism toward the predominance of pro-oxidant processes [28]. Glutathione system imbalance may also indicate activation of LPO reactions [29]. CP, characterized as a protein involved in metabolism of free radicals in extracellular compartments, is a copper-binding glycoprotein. The antioxidant function of CP may also be related to its ability to bind copper ions, which are well-known pro-oxidant catalysts [30]. Thus, severe insufficiency of these components may negatively affect antioxidant status in patients with MJ.

    Under conditions of bile duct decompression in MJ, increased oxygenation of liver tissues occurs, which contributes to the phenomenon of “oxygen paradox”, when the supply of the usual amount of oxygen to ischemic tissues, where pro-oxidants predominate over antioxidants, leads to additional formation of reactive oxygen species [8]. These conditions lead to further cell membrane lipid layer damage and changes to physical and chemical properties, followed by inflammation activation at both local and systemic levels [31]. Potential mechanisms of activation of oxidative stress in the intestinal region in MJ include: increased levels of bile acids, systemic endotoxemia, inducible nitric oxide synthase increased expression [3],[32], increased neutrophil chemotaxis and generation of superoxide anions [33], and antioxidant component systemic level decreases [8].

    Surgical intervention improved LPO-AOD scores in the MJNT group in comparison with both the control group and with preoperative data. Thus, after surgical intervention, patients with MJNT showed a decrease in TBARs and an increase in catalase and CP levels. In the post-surgery MJT group, the parameters of the LPO-AOD system also changed in comparison with the control group and with pre-surgery data. In this group, there was a decrease in the primary products of LPO and CDs, and an increase in the level of catalase and CP relative to pre-surgery data. Note that surgery can provoke development of septic complications post-surgery [3],[26],[34]. In addition, in the postoperative period, progression of liver failure and endogenous intoxication is noted; this is the main cause of death in MJ [1],[4]. Patients with MJ in the postoperative period show decreased humoral immunity parameters, development of leukocytosis, and decreased phagocytic activity [26]. There is disturbance of the functional activity of macrophages and granulocytes, and expression of HLA-DR antigens decreases rapidly on blood monocytes after surgery [35]. All classes immunoglobulins, especially IgG, increase, along with the number of T-suppressors, suppressor macrophages, synthesis of interleukins-4,-10, transforming growth factor, and prostaglandin E; in addition, the function of natural killer cells is inhibited and the number of T helper cells decreases [35],[36]. Due to the numerous disorders of the immune response in MJ, detoxification of the body plays an important role in ensuring a favorable postoperative outcome. One detoxification measure is inactivation of toxic products of free radical reactions, since they play a decisive role in the functioning of the immune system [37]. It is known that the damaging effect of free radicals is realized both as a result of their direct effect on proteins, enzymes, and nucleic acids, and through LPO products (lipid hydroperoxides, conjugated dienes, aldehydes, formed at different stages of the chain reaction [38]. The most toxic final products are TBARs, which can inhibit DNA synthesis, induce apoptosis and thereby suppress the proliferation, maturation, and growth of body cells; TBARs also have a strong carcinogenic effect and cause mutational changes [39]. An increase in toxic products of LPO, as a rule, indicates rapid involvement of LPO processes in the pathogenetic mechanisms of structural and functional disorders in the cells of organs and tissues. Thus, a decrease in the concentration of TBARs products in patients with MJNT may serve as a favorable sign in the postoperative period, unlike the MJT group, where this change is not observed.

    5.   Conclusion

    This study shows high activity of lipid peroxidation processes accompanied by decreased antioxidant enzyme concentration in men with MJ, regardless of disease origin. At the same time, the presence of malignant growth in patients with MJ is characterized by more pronounced changes in lipid peroxidation-antioxidant defense system. Post-surgery, patients with MJNT experienced a decrease in final products; in patients with MJT, the level of primary products decreased, while catalase and CP increased in both groups relative to pre-surgery data. Based on the data obtained, it is advisable to include antioxidant drugs in treatment of patients with MJ, with special attention given to the tumor origin of the disease.



    [1] P. Li, Y. Zhang, L. Yuan, H. X. Xiao, B. B. Lin, X. H. Xu, Efficient long-short temporal attention network for unsupervised video object segmentation, Pattern Recogn., 146 (2024), 110078. https://doi.org/10.1016/j.patcog.2023.110078 doi: 10.1016/j.patcog.2023.110078
    [2] E. Moen, D. Bannon, T. Kudo, W. Graf, M. Covert, D. Van Valen, Deep learning for cellular image analysis, Nat. Methods, 16 (2019), 1233–1246. https://doi.org/10.1038/s41592-019-0403-1 doi: 10.1038/s41592-019-0403-1
    [3] L. Chena, P. Bentley, K. Mori, K. Misawa, M. Fujiwara, D. Rueckert, Self-supervised learning for medical image analysis using image context restoration, Med. Image Anal., 58 (2019). https://doi.org/10.1016/j.media.2019.101539 doi: 10.1016/j.media.2019.101539
    [4] Y. A. Nanehkaran, D. F. Zhang, J. D. Chen, Y. Tian, N. Al-Nabhan, Recognition of plant leaf diseases based on computer vision, J. Ambient Intell. Human. Comput., (2020), 1–18. https://doi.org/10.1007/s12652-020-02505-x doi: 10.1007/s12652-020-02505-x
    [5] M. Wankhade, A. C. S. Rao, C. Kulkarni, A survey on sentiment analysis methods, applications, and challenges, Artif. Intell. Rev., 55 (2022), 5731–5780. https://doi.org/10.1007/s10462-022-10144-1 doi: 10.1007/s10462-022-10144-1
    [6] D. M. E. D. M. Hussein, A survey on sentiment analysis challenges, J. King Saud Univ. Eng. Sci., 30 (2018), 330–338. https://doi.org/10.1016/j.jksues.2016.04.002 doi: 10.1016/j.jksues.2016.04.002
    [7] K. R. Chowdhary, Natural language processing, in Fundamentals of Artificial Intelligence, Springer, (2020), 603–649. https://doi.org/10.1007/978-81-322-3972-7_19
    [8] V. Raina, S. Krishnamurthy, Natural language processing, in Building an Effective Data Science Practice, Springer, (2022), 63–73. https://doi.org/10.1007/978-1-4842-7419-4_6
    [9] M. Malik, M. K. Malik, K. Mehmood, I. Makhdoom, Automatic speech recognition: A survey, Multimed. Tools Appl., 80 (2021), 9411—9457. https://doi.org/10.1007/s11042-020-10073-7 doi: 10.1007/s11042-020-10073-7
    [10] D. Wang, X. D. Wang, S. H. Lv, An overview of end-to-end automatic speech recognition, Symmetry, 11 (2019), 1018. https://doi.org/10.3390/sym11081018 doi: 10.3390/sym11081018
    [11] L. Deng, X. Li, Machine learning paradigms for speech recognition: An overview, IEEE Trans. Audio, 21 (2013), 1060–1089. https://doi.org/10.1109/TASL.2013.2244083 doi: 10.1109/TASL.2013.2244083
    [12] X. Tan, T. Qin, F. Soong, T. Y. Liu, A survey on neural speech synthes, preprint, arXiv: 2106.15561.
    [13] V. Mario, G. Angiulli, P. Crucitti, D. D. Carlo, F. Laganà, D. Pellicanò, et al., A fuzzy similarity-based approach to classify numerically simulated and experimentally detected carbon fiber-reinforced polymer plate defects, Sensors, 22 (2022), 4232. https://doi.org/10.3390/s22114232 doi: 10.3390/s22114232
    [14] M. Versaci, G. Angiulli, P. D. Barba, F. C. Morabito, Joint use of eddy current imaging and fuzzy similarities to assess the integrity of steel plates, Open Phys., 18 (1) (2020), 230–240. https://doi.org/10.1515/phys-2020-0159 doi: 10.1515/phys-2020-0159
    [15] W. Zeng, H. L. Zhu, C. Lin, Z. Y. Xiao, A survey of generative adversarial networks and their application in text-to-image synthesis, Elect. Res. Arch., 31 (2023), 7142–7181. https://doi.org/10.3934/era.2023362 doi: 10.3934/era.2023362
    [16] I. Goodfellow, P. A. Jean, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, et al., Generative adversarial nets, in 2014 Advances in Neural Information Processing Systems (NIPS), 27 (2014), 1–9.
    [17] T. Y. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, et al., Microsoft COCO: Common objects in context, in 2014 European conference computer vision (ECCV), (2014), 740–755. https://doi.org/10.1007/978-3-319-10602-1_48
    [18] J. Zou, M. Huss, A. Abid, P. Mohammadi, A. Torkamani, A. Telenti, A primer on deep learning in genomics, Nat. Genet., 51 (2019), 12–18. https://doi.org/10.1038/s41588-018-0295-5 doi: 10.1038/s41588-018-0295-5
    [19] A. Borji, S. Izadi, L. Itti, iLab-20M: A large-scale controlled object dataset to investigate deep learning, in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), (2016), 2221–2230. https://doi.org/10.1109/CVPR.2016.244
    [20] O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, et al., ImageNet large scale visual recognition challenge, Int. J. Comput. Vis., 115 (2015), 211–252. https://doi.org/10.1007/s11263-015-0816-y doi: 10.1007/s11263-015-0816-y
    [21] K. M. He, X. Y. Zhang, S. Q. Ren, J. Sun, Deep residual learning for image recognition, in 2016 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2016), 770–778. https://doi.org/10.1109/CVPR.2016.90
    [22] A. G. Howard, M. L. Zhu, B. Chen, D. Kalenichenko, W. J. Wang, T. Weyand, et al., MobileNets: Efficient convolutional neural networks for mobile vision applications, preprint, arXiv: 1704.04861.
    [23] X. Y. Zhang, X. Y. Zhou, M. X. Lin, J. Sun, ShuffleNet: An extremely efficient convolutional neural network for mobile devices, in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2018), 6848–6856. https://doi.org/10.1109/CVPR.2018.00716
    [24] W. Zeng, Z. Y. Xiao, Few-shot learning based on deep learning: A survey, Math. Biosci. Eng., 21 (2024), 679–711. https://doi.org/10.3934/mbe.2024029 doi: 10.3934/mbe.2024029
    [25] J. Yang, X. M. Wang, Z. P. Luo, Few-shot remaining useful life prediction based on meta-learning with deep sparse kernel network, Inform. Sci., 653 (2024), 119795. https://doi.org/10.1016/j.ins.2023.119795 doi: 10.1016/j.ins.2023.119795
    [26] Y. Q. Wang, Q. M. Yao, J. T. Kwok, L. M. Ni, Generalizing from a few examples: A survey on few-shot learning, ACM Comput. Surveys, 53 (2020), 1–34. https://doi.org/10.1145/3386252 doi: 10.1145/3386252
    [27] C. Shorten, T. M. Khoshgoftaar, A survey on Image Data Augmentation for Deep Learning, J. Big Data, 6 (2019), 60. https://doi.org/10.1186/s40537-019-0197-0 doi: 10.1186/s40537-019-0197-0
    [28] N. E. Khalifa, M. Loey, S. Mirjalili, A comprehensive survey of recent trends in deep learning for digital images augmentation, Artif. Intell. Rev., 55 (2022), 2351-–2377. https://doi.org/10.1007/s10462-021-10066-4 doi: 10.1007/s10462-021-10066-4
    [29] K. Alomar, H. I. Aysel, X. H. Cai, Data augmentation in classification and segmentation: A survey and new strategies, J. Imaging, 9 (2023), 46. https://doi.org/10.3390/jimaging9020046 doi: 10.3390/jimaging9020046
    [30] T. DeVries, G. W. Taylor, Improved regularization of convolutional neural networks with cutout, preprint, arXiv: 1708.04552.
    [31] N. H. Li, S. J. Liu, Y. Q. Liu, S. Zhao, M. Liu, Random erasing data augmentation, in Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 34 (2020), 13001–13008. https://doi.org/10.1609/aaai.v34i07.7000
    [32] K. K. Singh, Y. J. Lee, Hide-and-Seek: Forcing a network to be meticulous for weakly-supervised object and action localization, in 2017 IEEE International Conference on Computer Vision (ICCV), IEEE, (2017), 3544–3553. https://doi.org/10.1109/ICCV.2017.381
    [33] P. G. Chen, S. Liu, H. S. Zhao, X. G. Wang, J. Y. Jia, GridMask data augmentation, preprint, arXiv: 2001.04086.
    [34] E. D. Cubuk, B. Zoph, D. Mane, V. Vasudevan, Q. V. Le, AutoAugment: Learning augmentation policies from data, preprint, arXiv: 1805.09501.
    [35] S. Lim, I. Kim, T. Kim, C. Kim, S. Kim, Fast autoaugment, in 2019 Advances in Neural Information Processing Systems (NIPS), (2019).
    [36] R. Hataya, J. Zdenek, K. Yoshizoe, H. Nakayama, Faster autoaugment: Learning augmentation strategies using backpropagation, in 2020 European conference computer vision (ECCV), (2022), 1–16. https://doi.org/10.1007/978-3-030-58595-2_1
    [37] E. D. Cubuk, B. Zoph, J. Shlens, Q. V. Le, Faster autoaugment: Learning augmentation strategies using backpropagation, in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), (2020), 3008–3017. https://doi.org/10.1109/CVPRW50498.2020.00359
    [38] D. Hendrycks, N. Mu, E. D. Cubuk, B. Zoph, J. Gilmer, B. Lakshminarayanan, Augmix: A simple data processing method to improve robustness and uncertainty, preprint, arXiv: 1912.02781.
    [39] K. Baek, D. Bang, H. Shim, GridMix: Strong regularization through local context mapping, Pattern Recogn., 109 (2021), 107594. https://doi.org/10.1016/j.patcog.2020.107594 doi: 10.1016/j.patcog.2020.107594
    [40] S. Yun, D. Han, S. Chun, S. J. Oh, S. Chun, J. Choe, et al., CutMix: Regularization strategy to train strong classifiers with localizable features, in 2019 IEEE/CVF International Conference on Computer Vision (ICCV), IEEE, (2019), 6022–6031. https://doi.org/10.1109/ICCV.2019.00612
    [41] M. Hong, J. Choi, G. Kim, StyleMix: Separating content and style for enhanced data augmentation, in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2021), 14857–14865. https://doi.org/10.1109/CVPR46437.2021.01462
    [42] D. Walawalkar, Z. Q. Shen, Z. C. Liu, M. Savvides, Attentive cutmix: An enhanced data augmentation approach for deep learning based image classification, preprint, arXiv: 2003.13048.
    [43] H. Y. Zhang, M. Cisse, Y. N. Dauphin, D. Lopez-Paz, Mixup: Beyond empirical risk minimization, preprint, arXiv: 1710.09412.
    [44] E. Harris, A. Marcu, M. Painter, M. Niranjan, A. Prügel-Bennett, J. Hare, Fmix: Enhancing mixed sample data augmentation, preprint, arXiv: 2002.12047.
    [45] J. Qin, J. M. Fang, Q. Zhang, W. Y. Liu, X. G. Wang, X. G. Wang, Resizemix: Mixing data with preserved object information and true labels, preprint, arXiv: 2012.11101.
    [46] A. F. M. S. Uddin, M. S Monira, W. Shin, T. C. Chung, S. H. Bae, Saliencymix: A saliency guided data augmentation strategy for better regularization, preprint, arXiv: 2006.01791.
    [47] A. Bochkovskiy, C. Y. Wang, H. Y. M. Liao, Yolov4: Optimal speed and accuracy of object detection, preprint, arXiv: 2004.10934.
    [48] J. H. Liu, B. X. Liu, H. Zhou, H. S. Li, Y. Liu, Tokenmix: Rethinking image mixing for data augmentation in vision transformers, in 2022 European conference computer vision (ECCV), (2022), 455–471. https://doi.org/10.1007/978-3-031-19809-0_26
    [49] M. Z. Chen, M. B. Lin, Z. H. Lin, Y. X. Zhang, F. Chao, R. R. Ji, SMMix: Self-Motivated Image Mixing for Vision Transformers, in 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2023), 17214–17224. https://doi.org/10.1109/ICCV51070.2023.01583
    [50] L. F. Yang, X. Li, B. Zhao, R. J. Song, J. Yang, RecursiveMix: Mixed Learning with History, in 2020 Advances in Neural Information Processing Systems (NIPS), (2022), 8427–8440.
    [51] V. Verma, A. Lamb, C. Beckham, A. Najafi, I. Mitliagkas, D. Lopez-Paz, et al., Manifold mixup: Better representations by interpolating hidden states., in Proceedings of the 36th International Conference on Machine Learning (ICML), 97 (2019), 6438–6447.
    [52] J. H. Kim, W. Choo, H. Jeong, H. O. Song, Co-mixup: Saliency guided joint mixup with supermodular diversity, preprint, arXiv: 2102.03065.
    [53] J. H. Kim, W. Choo, H. O. Song, Puzzle mix: Exploiting saliency and local statistics for optimal mixup, in Proceedings of the 37th International Conference on Machine Learning (ICML), 119 (2020), 5275–5285.
    [54] A. Dabouei, S. Soleymani, F. Taherkhani, N. M. Nasrabadi, SuperMix: Supervising the mixing data augmentation, in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2021), 13789–13798. https://doi.org/10.1109/CVPR46437.2021.01358
    [55] C. Y. Gong, D. L. Wang, M. Li, V. Chandra, Q. Liu, KeepAugment: A simple information-preserving data augmentation approach, in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2021), 1055–1064. https://doi.org/10.1109/CVPR46437.2021.00111
    [56] M. Kang, S. Kim, GuidedMixup: An efficient mixup strategy guided by saliency maps, in 2023 Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), (2023), 1096–1104. https://doi.org/10.1609/aaai.v37i1.25191
    [57] T. Hong, Y. Wang, X. W. Sun, F. Z. Lian, Z. H. Kang, J. W. Ma, GradSalMix: Gradient saliency-based mix for image data augmentation, in 2023 IEEE International Conference on Multimedia and Expo (ICME), IEEE, (2023), 1799–1804. https://doi.org/10.1109/ICME55011.2023.00309
    [58] M. Mirza, S. Osindero, Conditional generative adversarial nets, preprint, arXiv: 1411.1784v1.
    [59] A. Odena, C. Olah, J. Shlens, Conditional image synthesis with auxiliary classifier GANs, in 2017 Proceedings of the 34rd International Conference on International Conference on Machine Learning, PMLR, (2017), 2642–2651.
    [60] G. Douzas, F. Bacao, Effective data generation for imbalanced learning using conditional generative adversarial networks, Expert Syst. Appl., 91, (2018), 464–471. https://doi.org/10.1016/j.eswa.2017.09.030
    [61] A. Antoniou, A. Storkey, H. Edwards, Data augmentation generative adversarial networks, preprint, arXiv: 1711.04340.
    [62] G. Mariani, F. Scheidegger, R. Istrate, C. Bekas, C. Malossi, Bagan: Data augmentation with balancing gan, preprint, arXiv: 1803.09655.
    [63] S. W. Huang, C. T. Lin, S. P. Chen, Y. Y. Wu, P. H. Hsu, S. H. Lai, Auggan: Cross domain adaptation with gan-based data augmentation, in 2018 Proceedings of the European Conference on Computer Vision (ECCV), (2018), 731—744. https://doi.org/10.1007/978-3-030-01240-3_44
    [64] X. Y. Zhu, Y. F. Liu, J. H. Li, T. Wan, Z. H. Qin, Emotion classification with data augmentation using generative adversarial networks, in 2018 Advances in Knowledge Discovery and Data Mining (PAKDD), 10939 (2018), 349—360. https://doi.org/10.1007/978-3-319-93040-4_28
    [65] E. Schwartz, L. Karlinsky, J. Shtok, S. Harary, M. Marder, A. Kumar, et al., Delta-encoder: An effective sample synthesis method for few-shot object recognition, in 2018 Advances in Neural Information Processing Systems (NIPS), 31 (2018).
    [66] A. Ali-Gombe, E. Elyan, MFC-GAN: Class-imbalanced dataset classification using multiple fake class generative adversarial network, Neurocomputing, 361 (2019), 212–221. https://doi.org/10.1016/j.neucom.2019.06.043 doi: 10.1016/j.neucom.2019.06.043
    [67] H. Yang, Y. Zhou, Ida-gan: A novel imbalanced data augmentation gan, in 2020 International Conference on Pattern Recognition (ICPR), IEEE, (2020), 8299-8305. https://doi.org/10.1109/ICPR48806.2021.9411996
    [68] A. Krizhevsky, Learning Multiple Layers of Features from Tiny Images, Master's thesis, University of Toronto, 2009.
    [69] J. Deng, W. Dong, R. Socher, L. J. Li, K. Li, F. F. Li, ImageNet: A large-scale hierarchical image database, in 2009 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2009), 248–255. https://doi.org/10.1109/CVPR.2009.5206848
    [70] D. Han, J. Kim, J. Kim, Deep pyramidal residual networks, in 2017 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2017), 6307–6315. https://doi.org/10.1109/CVPR.2017.668
    [71] A. Mikołajczyk, M. Grochowski, Data augmentation for improving deep learning in image classification problem, in 2018 International Interdisciplinary PhD Workshop (IIPhDW), IEEE, (2018), 117–122. https://doi.org/10.1109/IIPHDW.2018.8388338
    [72] C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, Z. Wojna, Rethinking the inception architecture for computer vision, in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2016), 2818–2826. https://doi.org/10.1109/CVPR.2016.308
    [73] G. Ghiasi, T. Y. Lin, Q. V. Le, Dropblock: A regularization method for convolutional networks, in 2018 Advances in Neural Information Processing Systems (NIPS), 31 (2018).
    [74] G. Huang, Y. Sun, Z. Liu, D. Sedra, K. Q. Weinberger, Deep networks with stochastic depth, in 2016 European Conference Computer Vision (ECCV), (2016), 646–661. https://doi.org/10.1007/978-3-319-46493-0_39
    [75] J. J. Bird, C. M. Barnes, L. J. Manso, A. Ekárt, D. R. Faria, Fruit quality and defect image classification with conditional GAN data augmentation, Sci. Hortic., 293 (2022), 110684. https://doi.org/10.1016/j.scienta.2021.110684 doi: 10.1016/j.scienta.2021.110684
    [76] H. M. Gao, J. P. Zhang, X. Y. Cao, Z. H. Chen, Y. Y. Zhang, C. M. Li, Dynamic data augmentation method for hyperspectral image classification based on Siamese structure, J. Sel. Top. Appl. Earth Observ. Remote Sens., 14 (2021), 8063–8076. https://doi.org/10.1109/JSTARS.2021.3102610 doi: 10.1109/JSTARS.2021.3102610
    [77] O. A. Shawky, A. Hagag, E. S. A. El-Dahshan, M. A. Ismail, Remote sensing image scene classification using CNN-MLP with data augmentation, Optik, 221 (2020), 165356. https://doi.org/10.1016/j.ijleo.2020.165356 doi: 10.1016/j.ijleo.2020.165356
    [78] O. O. Abayomi-Alli, R. Damaševičius, S. Misra, R. Maskeliūnas, Cassava disease recognition from low-quality images using enhanced data augmentation model and deep learning, Expert Syst., 38 (2021), e12746. https://doi.org/10.1111/exsy.12746 doi: 10.1111/exsy.12746
    [79] Q. H. Cap, H. Uga, S. Kagiwada, H. Iyatomi, Leafgan: An effective data augmentation method for practical plant disease diagnosis, IEEE Trans. Autom. Sci. Eng., 19 (2022), 1258–1267. https://doi.org/10.1109/TASE.2020.3041499 doi: 10.1109/TASE.2020.3041499
    [80] W. Li, C. C. Gu, J. L. Chen, C. Ma, X. W. Zhang, B. Chen, et al., DLS-GAN: Generative adversarial nets for defect location sensitive data augmentation, IEEE Trans. Autom. Sci. Eng., (2023), 1–17. https://doi.org/10.1109/TASE.2023.3309629
    [81] S. Jain, G. Seth, A. Paruthi, U. Soni, G. Kumar, Synthetic data augmentation for surface defect detection and classification using deep learning, J. Intell. Manuf., 33 (2022), 1007–1020. https://doi.org/10.1007/s10845-020-01710-x doi: 10.1007/s10845-020-01710-x
    [82] Y. L. Wang, G. Huang, S. J. Song, X. R. Pan, Y. T. Xia, C. Wu, Regularizing deep networks with semantic data augmentation, IEEE Trans. Pattern Anal. Mach. Intell., 44 (2021), 3733–3748. https://doi.org/10.1109/TPAMI.2021.3052951 doi: 10.1109/TPAMI.2021.3052951
    [83] B. Zoph, E. D. Cubuk, G. Ghiasi, T. Y. Lin, J. Shlens, Q. V. Le, Learning data augmentation strategies for object detection, in 2020 Proceedings of the European Conference on Computer Vision (ECCV), 12372 (2020), 566—583. https://doi.org/10.1007/978-3-030-58583-9_34
    [84] Y. Tang, B. P. Li, M. Liu, B. Y. Chen, Y. N. Wang, W. L. Ouyang, Autopedestrian: An automatic data augmentation and loss function search scheme for pedestrian detection, IEEE Trans. Image Process., 30 (2021), 8483–8496. https://doi.org/10.1109/TIP.2021.3115672 doi: 10.1109/TIP.2021.3115672
    [85] C. L. Wang, Z. F Xiao, Lychee surface defect detection based on deep convolutional neural networks with gan-based data augmentation, Agronomy, 11 (2021), 1500. https://doi.org/10.3390/agronomy11081500 doi: 10.3390/agronomy11081500
    [86] W. W. Zhang, Z. Wang, C. C. Loy, Exploring data augmentation for multi-modality 3D object detection, preprint, arXiv: 2012.12741.
    [87] C. W. Wang, C. Ma, M. Zhu, X. K. Yang, Pointaugmenting: Cross-modal augmentation for 3D object detection, in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (2021), 11789–11798. https://doi.org/10.1109/CVPR46437.2021.01162
    [88] Y. W. Li, A. W. Yu, T. J. Meng, B. Caine, J. Ngiam, D. Y. Peng, et al., Deepfusion: Lidar-camera deep fusion for multi-modal 3d object detection, in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (2022), 17161–17170. https://doi.org/10.1109/CVPR52688.2022.01667
    [89] S. Y. Cheng, Z. Q. Leng, E. D. Cubuk, B. Zoph, C. Y. Bai, J. Ngiam, et al., Improving 3d object detection through progressive population based augmentation, in 2020 Proceedings of the European Conference on Computer Vision (ECCV), 12366 (2020), 279–294. https://doi.org/10.1007/978-3-030-58589-1_17
    [90] X. K. Zhu, S. C. Lyu, X. Wang, Q. Zhao, TPH-YOLOv5: Improved YOLOv5 based on transformer prediction head for object detection on drone-captured scenarios, in 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), (2021), 2778–2788. https://doi.org/10.1109/ICCVW54120.2021.00312
    [91] X. M. Sun, X. C. Jia, Y. Q. Liang, M. G. Wang, X. B. Chi, A defect detection method for a boiler inner wall based on an improved YOLO-v5 network and data augmentation technologies, IEEE Access, 10 (2022), 93845–93853. https://doi.org/10.1109/ACCESS.2022.3204683 doi: 10.1109/ACCESS.2022.3204683
    [92] W. Y. Liu, G. F. Ren, R. S. Yu, S. Guo, J. K. Zhu, L. Zhang, Image-adaptive YOLO for object detection in adverse weather conditions, in 2022 Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 36 (2022), 1792–1800. https://doi.org/10.1609/aaai.v36i2.20072
    [93] Q. M. Chung, T. D. Le, T. V. Dang, N. D. Vo, T. V. Nguyen, K. Nguyen, Data augmentation analysis in vehicle detection from aerial videos, in 2020 RIVF International Conference on Computing and Communication Technologies (RIVF), (2022), 1–3. https://doi.org/10.1109/RIVF48685.2020.9140740
    [94] D. Su, H. Kong, Y. L. Qiao, S. Sukkarieh, Data augmentation for deep learning based semantic segmentation and crop-weed classification in agricultural robotics, Comput. Electron. Agric., 190 (2021), 106418. https://doi.org/10.1016/j.compag.2021.106418 doi: 10.1016/j.compag.2021.106418
    [95] J. Choi, T. Kim, C. Kim, Self-ensembling with gan-based data augmentation for domain adaptation in semantic segmentation, in 2019 IEEE/CVF International Conference on Computer Vision (ICCV), IEEE, (2019), 6829–6839. https://doi.org/10.1109/ICCV.2019.00693
    [96] J. L. Yuan, Y. F. Liu, C. H. Shen, Z. B. Wang, H. Li, A simple baseline for semi-supervised semantic segmentation with strong data augmentation, in 2021 IEEE/CVF International Conference on Computer Vision (ICCV), IEEE, (2021), 8209–8218. https://doi.org/10.1109/ICCV48922.2021.00812
    [97] S. T. Liu, J. Q. Zhang, Y. X. Chen, Y. F. Liu, Z. C. Qin, T. Wan, Pixel level data augmentation for semantic image segmentation using generative adversarial networks, in 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, (2019), 1902–1906. https://doi.org/10.1109/ICASSP.2019.8683590
    [98] I. Budvytis, P. Sauer, T. Roddick, K. Breen, R. Cipolla, Large scale labelled video data augmentation for semantic segmentation in driving scenarios, in 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), IEEE, (2017), 230–237. https://doi.org/10.1109/ICCVW.2017.36
    [99] V. Olsson, W. Tranheden, J. Pinto, L. Svensson, Classmix: Segmentation-based data augmentation for semi-supervised learning, in 2021 IEEE Winter Conference on Applications of Computer Vision (WACV), IEEE, (2021), 1368–1377. https://doi.org/10.1109/WACV48630.2021.00141
    [100] J. W. Zhang, Y. C. Zhang, X. W. Xu, Objectaug: Object-level data augmentation for semantic image segmentation, in 2021 International Joint Conference on Neural Networks (IJCNN), IEEE, (2021), 1–8. https://doi.org/10.1109/IJCNN52387.2021.9534020
  • Reader Comments
  • © 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索
Mathematical Biosciences and Engineering
4.4

Metrics

Article views(3996) PDF downloads(255) Cited by(10)

Preview PDF
Download XML
Export Citation
Article outline
Abstract
   Introduction
   Materials and methods
   Results
   Discussion
   Conclusion
References

Figures and Tables

Figures(18)  /  Tables(4)

Other Articles By Authors

Related pages

Tools

Copyright © AIMS Press

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog