测试自建训练集ann识别
识别自己的字，识别率和层数变化如下：
| 层数 | 最高对 | 次高对 |
| 2 | 41.72 | |
| 3 | 68.10 | |
| 4 | 85.69 | 3.28 |
| 5 | 92.76 | 1.55 |
| 6 | 93.79 | 0.69 |
| 7 | 97.07 | 0.86 |
| 8 | 96.55 | 0.86 |
| 9 | 97.07 | 0.52 |
| 10 | 97.93 | 1.21 |
| 11 | 97.41 | 0.69 |
更改训练集和测试集文件，数据基本变化不大，即10层时均在95%以上。
识别女士写的字，识别率80%
结果如下：
| 层数 | 最高对 | 次高对 |
| 7 | 77.07 | 5.69 |
| 8 | 79.83 | 3.97 |
| 9 | 78.10 | 4.48 |
| 10 | 81.21 | 4.14 |
| 11 | 82.93 | 3.97 |
| 12 | 82.76 | 4.31 |
| 13 | 80.52 | 4.48 |
女士测试集如下：
主要错误如下
77.07 5.69 layer=7
(1 3), (1 10),
(2 4),
(3 0), (3 1), (3 6), (3 8), (3 9), (3 14), (3 16), (3 17), (3 18), (3 19), (3 21), (3 24),
(4 9), (4 11), (4 16), (4 24), (4 27), (4 28),
(5 0), (5 8), (5 12), (5 13), (5 15), (5 20), (5 22), (5 26), (5 27), (5 28),
(6 3), (6 4), (6 6), (6 20), (6 24),
(8 4), (8 8), (8 16), (8 17), (8 23),
(9 0), (9 1), (9 2), (9 3), (9 5), (9 6), (9 7), (9 13), (9 14), (9 19), (9 25),
(10 21), (10 26),
(11 5), (11 7), (11 10), (11 14), (11 21), (11 27), (11 28),
(12 3), (12 4), (12 10), (12 11), (12 12), (12 24),
(13 1), (13 10), (13 16), (13 17), (13 22), (13 23), (13 24),
(14 2), (14 7), (14 10), (14 13), (14 15), (14 16), (14 19), (14 20), (14 23), (14 24), (14 25), (14 28),
(15 1), (15 3), (15 4), (15 5), (15 6), (15 7), (15 8), (15 9), (15 10), (15 11), (15 12), (15 14), (15 15), (15 16), (15 18), (15 20), (15 22), (15 24), (15 28),
(17 8), (17 9), (17 13), (17 18), (17 24), (17 25),
(18 2), (18 8), (18 9), (18 10), (18 12), (18 14), (18 15), (18 17), (18 19), (18 20), (18 21), (18 23), (18 27),
(19 5), (19 14), (19 16), (19 21), (19 24), (19 25), (19 26), (19 27), (19 28),
下一步努力方向，找到解决人差异性手写体数字的识别问题。
