Automatic Generation of Character Animation using Forced Switching Mechanism based Arithmetic Optimization Algorithm with Convolutional Neural Network

Nowadays in China, the character animation industry is new and developing industry which appears in world stage frequently and widely involved and respected by people from the world. Therefore, this research proposes a Forced Switching Mechanism based Arithmetic Optimization Algorithm (FSMAOA) with Convolutional Neural Network (CNN) for generating character animation. The FSMAOA is used for feature selection which is applied to escape local optima and enhancing the balance among global and local search. Then, the CNN is used for generating character animation which recognize and acquire the spatial and textural features of characters in images. The collected data is preprocessed by using min-max normalization which is used to standardize each image based on minimum and maximum values. The FSMAOACNN performance is measured by using metrics of accuracy, precision, sensitivity and f-measure. The FSMAOA-CNN achieves $\mathbf{9 7. 8 6 \%}$ accuracy, $\mathbf{9 6. 5 8 \%}$ precision, $\mathbf{9 5. 4 2 \%}$ sensitivity and $96.15 \% \mathrm{f}$-measure for the collected dataset.