No abstract

Many inscriptions suffer severe damage as a result of artificial or natural causes, necessitating computer repair and protection. There isn’t a publicly accessible dataset of inscriptions, and the majority of character inpainting algorithms that are currently in use are directly derived from common image inpainting algorithms , which have insufficient feature extraction capabilities for the character of inscriptions and don’t have metrics for the weights of obscured and unobscured regions. We developed a Character Auto-Encoder (CAE) to improve the inpaint-ing capabilities of inscription characters and solve the aforementioned issues. The down-sampling module is replaced by the Branch Convolutional Channel Attention Module (BCCAM), which employs a branching structure to enhance the model’s capacity for representation and weights various areas of the character. First, we compared other inpainting models using the inscription dataset, and the CAE’s peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) increased by 0.82 and 0.016, respectively, compared to the baseline model. Next, we compared the Handwritten Chinese Character Inpainting Generative Adver-sarial Network (HCCI-GAN) with the handwritten Chinese character dataset, and the CAE’s design was relatively more ingenious and simpler.

Many inscriptions have suffered substantial damage as a result of artificial or natural causes, necessitating the use of computers for their repair and preservation. The character reconstruction method, however, is not sufficiently improved by the character inpainting models now in use. The traditional deconvolutional up-sampling module can only extract local spatial points from low-resolution feature maps to produce high-resolution details. To solve this problem, we design Binary Auto-Encoder (BAE) by introducing the self-attention mechanism in the U-Net decoder. The self-attention module calculates the feature weighting at each place to determine responsiveness and thereby computes the attention vector at a very cheap cost. Additionally, we use a spectrum normalization strategy to enhance the decoder’s up-sampling procedure. The peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) of BAE improve by 1.9 and 0.023, respectively, when compared to the baseline model, according to the results of our pairwise comparison of various inpainting models using the binary inscription dataset. The Handwritten Chinese Character Inpainting Generative Adversarial Network (HCCI-GAN) is then compared using the dataset of handwritten Chi-nese characters, and the findings demonstrate that the BAE has a cleverer and simpler architecture.