Research Article
Pattern Recognition of Holographic Image Library Based on Deep Learning
Table 6
Function test cases and qualification criteria.
| Function name | Test case | Expected output | Eligibility criteria |
| getLightCorrect | Color ID card image in highlight area | Correct the image | Visually distinguish the corrected image of the highlight area | ImageDiv_ Otsu_ Bersen | ID card image after grayscale | Binarized image | The characters are clearly separated from most of the background | getLineDiv | Binarized image | Character row and column split data | Split characters correctly | getGridDiv | Unit number normalization graph | Character elastic grid segmentation results | The projections within the grid are roughly equal | getHanziLine | Unit number normalization graph | The four fuzzy stroke features of Chinese characters | The four stroke characteristics of Chinese characters are correctly extracted through visual discrimination | getHanziFeature | Unit number normalization graph | Chinese character characteristics | No criterion |
|
|