基于眼部数据的挡车工表情识别方法 下载: 810次
Facial Expression Recognition Method of Spinner Based on Eye Data
图 & 表
图 2. 图像预处理过程
Fig. 2. Image preprocessing process
下载图片 查看原文
图 3. 眼部数据集
Fig. 3. Eye data set
下载图片 查看原文
图 4. Inception-v3网络结构
Fig. 4. Inception-v3 network structure
下载图片 查看原文
图 5. 精确度曲线
Fig. 5. Accuracy curve
下载图片 查看原文
图 8. 经过不同处理的图像识别精度的对比。(a)眼部截取;(b)图像预处理;(c)本文方法
Fig. 8. Comparison of image recognition accuracy after different processing. (a) Eye interception; (b) image preprocessing; (c) our method
下载图片 查看原文
表 1样本数据
Table1. Sample data
Samplenumber | Workingtime /h | Finenessunevenness /% | Number ofthin nodeper km | Number ofthick nodeper km | Number of3-mm hairinessper m | Breakingstrength /(10-2 N·tex-1) | Breakingelongation /% |
---|
1 | 1 | 21.79 | 756 | 219 | 2.34 | 5.91 | 8.77 | 1 | 2 | 21.74 | 756 | 219 | 2.64 | 6.03 | 8.78 | 1 | 3 | 21.89 | 763 | 229 | 2.7 | 5.92 | 8.79 | 1 | 4 | 21.87 | 763 | 210 | 1.98 | 5.94 | 8.79 | 1 | 5 | 21.97 | 767 | 218 | 3.8 | 6.04 | 8.88 | 1 | 6 | 22.08 | 758 | 217 | 4.38 | 6.15 | 8.67 | 1 | 7 | 21.95 | 764 | 213 | 5.47 | 6.05 | 8.84 | 1 | 8 | 21.94 | 765 | 219 | 2.41 | 6.07 | 8.87 | 2 | 1 | 21.97 | 745 | 210 | 1.96 | 5.98 | 8.90 | 2 | 2 | 21.81 | 752 | 220 | 2.03 | 6.12 | 8.76 | 2 | 3 | 21.95 | 748 | 215 | 2.41 | 5.96 | 8.82 | 2 | 4 | 21.92 | 752 | 220 | 2.35 | 6.1 | 8.91 | 2 | 5 | 22.05 | 758 | 217 | 2.36 | 5.89 | 8.75 | 2 | 6 | 21.87 | 756 | 215 | 1.86 | 5.94 | 8.75 | 2 | 7 | 21.94 | 756 | 218 | 2.04 | 5.99 | 8.64 | 2 | 8 | 22.08 | 756 | 221 | 2.41 | 6.11 | 8.77 | 3 | 1 | 22.05 | 752 | 216 | 2.33 | 5.96 | 8.94 | 3 | 2 | 22.06 | 761 | 223 | 3.14 | 5.88 | 8.64 | 3 | 3 | 21.97 | 760 | 224 | 2.85 | 6.01 | 8.59 | 3 | 4 | 21.86 | 750 | 220 | 2.53 | 5.94 | 8.88 | 3 | 5 | 22.07 | 752 | 224 | 2.63 | 6.08 | 8.75 | 3 | 6 | 21.96 | 764 | 218 | 1.78 | 5.84 | 8.83 | 3 | 7 | 21.94 | 762 | 225 | 2.67 | 5.92 | 8.79 | 3 | 8 | 22.01 | 756 | 218 | 3.12 | 6.04 | 8.79 |
|
查看原文
表 2挡车工的表情对成纱指标影响的权重比例
Table2. Weight proportion of yarn index influence for the expression of the spinner%
Sample number | Weight |
---|
Finenessunevenness | Number ofthin node | Number ofthick node | Number of3-mm hairiness | Breakingstrength | Breakingelongation |
---|
1 | 0 | 0 | 0 | 99 | 0 | 0 | 2 | 0 | 0 | 2 | 94 | 2 | 1 | 3 | 0 | 0 | 1 | 98 | 1 | 1 |
|
查看原文
邵景峰, 冯海强. 基于眼部数据的挡车工表情识别方法[J]. 激光与光电子学进展, 2020, 57(18): 181026. Jingfeng Shao, Haiqiang Feng. Facial Expression Recognition Method of Spinner Based on Eye Data[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181026.