Affected by the COVID-19 epidemic, students' learning progress has been affected in various aspect, and blended teaching has become a new option. Although the impact of the epidemic on teaching has been greatly reduced, the effect of pure online and offline teaching is not ideal. In this context, this paper mainly proposed improvements to the current calligraphy blended teaching model through deep learning. Personalized content recommendation is made according to the actual learning situation of students' course content, so as to strengthen students' understanding and cognition of calligraphy course content. Based on the deep learning theory, this paper constructed the IUNeu calligraphy course content recommendation model, and selects the third-grade students of school A as the experimental objects to conduct an experimental comparative analysis of different teaching modes. This paper analyzed from three perspectives of calligraphy theory course teaching, calligraphy practice course teaching and calligraphy ability. The experimental results were shown as follows. In the teaching of calligraphy theory courses, the pass rates of the students in the experimental group were 26.6%, 26.7%, 13.3%, 40%, and 40% higher than those in the control group in the origin of Chinese characters, the structure of Chinese characters, the strokes of Chinese characters, the physical evolution of Chinese characters, and the stippling of Chinese characters, respectively. In terms of calligraphy ability, the calligraphy ability of the students in the experimental group was higher than that of the control group, and the proportion of students in the four calligraphy appreciation abilities of connotation, artistic conception, temperament and distinguishing strokes was 26.6%, 26.7%, 33.4% and 13.3% higher than that of the control group. This showed that the effect of the new calligraphy course teaching mode based on deep learning is remarkable, and it also provided a certain experimental basis for the follow-up calligraphy course education reform.

Shuhang Li. Online and Offline Blended Teaching of Calligraphy Based on Deep Learning. International Journal of Educational Innovation and Science (2020), Vol. 1, Issue 1: 15-30. https://doi.org/10.38007/IJEIS.2020.010102.

[1] Wu Y, Wang J. Three-stage blended Chinese Teaching Online and Offline for International Students: A Case Study on Chinese Teaching for International Students in S University. Journal of Higher Education Research. (2022) 3(2): 207-211. https://doi.org/10.32629/jher.v3i2.758

[2] Wu X. Research on the Reform of Ideological and Political Teaching Evaluation Method of College English Course Based on "Online and Offline" Teaching. Journal of Higher Education Research. (2022) 3(1): 87-90. https://doi.org/10.32629/jher.v3i1.641

[3] Wang X, Zhang W. Improvement of Students' Autonomous Learning Behavior by Optimizing Foreign Language Blended Learning Mode. SAGE Open. (2022) 12(1): 28-38. https://doi.org/10.1177/21582440211071108

[4] Wang Y. Implications of Blended Teaching Based on Theory of Semantic Wave for Teaching English Writing in High School. Journal of Higher Education Research. (2022) 3(2): 166-168. https://doi.org/10.32629/jher.v3i2.747

[5] Ige O A, Hlalele D J. Effects of computer-aided and blended teaching strategies on students' achievement in civic education concepts in mountain learning ecologies. Education and Information Technologies. (2017) 22(6): 1-17. https://doi.org/10.1007/s10639-017-9598-x

[6] Doleck T, Lemay D J, Basnet R B. Predictive analytics in education: a comparison of deep learning frameworks. Education and Information Technologies. (2020) 25(3): 1-13. https://doi.org/10.1007/s10639-019-10068-4

[7] Li Q, Kumar P M, Alazab M. IoT-assisted physical education training network virtualization and resource management using a deep reinforcement learning system. Complex & Intelligent Systems. (2022) 8(2): 1229-1242. https://doi.org/10.1007/s40747-021-00584-7

[8] Tong M, Gao T, Peng J. Coastal soil pollution detection and business English teaching index construction based on deep feature fusion. Arabian Journal of Geosciences. (2021) 14(16): 1-14. https://doi.org/10.1007/s12517-021-07891-w

[9] Chen C, Polemis M, Stengos T. On the examination of non-linear relationship between market structure and performance in the US manufacturing industry. Economics Letters. (2018) 164(MAR.): 1-4. https://doi.org/10.1016/j.econlet.2017.12.030

[10] Zhai X, Ait-Si-Ali A, Amira A. MLP Neural Network Based Gas Classification System on Zynq SoC. IEEE Access. (2017) 4(99): 8138-8146. https://doi.org/10.1109/ACCESS.2016.2619181

[11] Shipsha A, Hallstrom S, Burman M. Effect of stacking sequence and bundle waviness in quasi-isotropic NCF composites subjected to compression. Composites. (2019) 178(Dec.1): 107423.1-107423.12. https://doi.org/10.1016/j.compositesb.2019.107423

[12] Pham G, Ebert K D. Diagnostic Accuracy of Sentence Repetition and Nonword Repetition for Developmental Language Disorder in Vietnamese. Journal of Speech Language and Hearing Research. (2020) 63(1): 1-16. https://doi.org/10.1044/2020_JSLHR-19-00366

[13] Yan J, Qi Y, Rao Q. LSTM-Based with Deterministic Negative Sampling for API Suggestion. International Journal of Software Engineering and Knowledge Engineering. (2019) 29(7): 1029-1051. https://doi.org/10.1142/S0218194019500347

[14] Li Q, Li H, Lu Z. Denoising of Hyperspectral Images Employing Two-Phase Matrix Decomposition. IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing. (2017) 7(9): 3742-3754. https://doi.org/10.1109/JSTARS.2014.2360409

[15] Cui S, Joe I. Collision prediction for a low power wide area network using deep learning methods. Journal of Communications and Networks. (2020) 22(3): 205-214. https://doi.org/10.1109/JCN.2020.000017

[16] Briand L, Bianculli D, Nejati S. The Case for Context-Driven Software Engineering Research: Generalizability Is Overrated. IEEE Software. (2017) 34(5): 72-75. https://doi.org/10.1109/MS.2017.3571562

[17] Takahashi K, Taki H, Tanabe S. An Automatic Occupation and Industry Coding System in Sociology. Journal of Natural Language Processing. (2017) 24(1): 135-170. https://doi.org/10.5715/jnlp.24.135

[18] Kang W, Wilcox L C. Mitigating the Curse of Dimensionality: Sparse Grid Characteristics Method for Optimal Feeback Control and HJB Equations. Computational Optimization & Applications. (2017) 68(11): 1-27. https://doi.org/10.1007/s10589-017-9910-0

[19] Belyi V, Gan W S. Integrated psychoacoustic active noise control and masking. Applied Acoustics. (2019) 145(FEB.):339-348. https://doi.org/10.1016/j.apacoust.2018.10.027

[20] Chigwada J P. Supporting Information Literacy Skills of Students for a Successful Transition to Higher Education: Opportunities and Challenges for Libraries in the Digital Era. International journal of digital library systems. (2019) 8(1):24-30. https://doi.org/10.4018/IJLIS.2019010102