基于红外热成像的早期疾病检测技术的研究进展 下载: 1281次
Research Progress of Early Disease Detection Technology Based on Infrared Thermography
1 贵州大学大数据与信息工程学院, 贵州 贵阳 550025
2 贵阳学院农产品无损检测工程研究中心, 贵州 贵阳, 550005
图 & 表
图 1. 红外热成像检测系统的工程流程
Fig. 1. Workflow of infrared thermal imaging detection system
下载图片 查看原文
图 2. 乳腺肿块的红外热像图[6]。(a)浸润性导管癌患者;(b)纤维腺瘤患者;(c)囊肿患者
Fig. 2. Infrared thermograms of breast masses[6]. (a)Patient with invasive ductal carcinoma; (b)patient with fibroadenoma; (c)patient with cyst
下载图片 查看原文
图 3. 皮肤癌检测算法框图
Fig. 3. Skin cancer detection algorithm block diagram
下载图片 查看原文
图 4. 腕管综合征患者手部的红外热像图对比[12]。(a)术前;(b)术后;(c)健康志愿者
Fig. 4. Comparison of infrared thermograms of hands in patients with carpal tunnel syndrome[12].(a)Before operation; (b)after operation;(c)healthy volunteer
下载图片 查看原文
图 5. 基于红外热成像数据的心肌缺血预诊断方法流程图[21]
Fig. 5. Process of myocardial ischemic pre-diagnosis based on infrared imaging data[21]
下载图片 查看原文
图 6. 面部红外热像图。(a)左侧周围性面瘫患者;(b)中枢性面瘫患者;(c)健康志愿者
Fig. 6. Infrared thermograms of face.(a)Patient with left-sided peripheral facial paralysis;(b)patient with central facial paralysis;(c)healthy volunteer
下载图片 查看原文
图 7. 卷积神经网络的基本结构图
Fig. 7. Basic structure diagram of convolutional neural network
下载图片 查看原文
表 1钼靶、超声和远红外热成像诊断直径<2 cm乳腺病灶效能的比较(n=853)[18]
Table1. Comparison of mammography, ultrasonography, and far-infrared thermography in the diagnosis of breast lesions less than 2 cm in diameter (n=853)[18]
Item | Pathological examination | Accuracy/% | Sensitivity/% | Specificity/% |
---|
Malignant | Benign |
---|
Mammography | | | | | | Malignant | 59 | 19 | 96.1 | 80.8 | 97.6 | Benign | 14 | 761 | Color Doppler ultrasonography | | | | | | Malignant | 63 | 26 | 95.8 | 86.3 | 96.6 | Benign | 10 | 754 | Far-infrared thermography | | | | | | Malignant | 66 | 17 | 97.1 | 90.4 | 97.8 | Benign | 7 | 763 |
|
查看原文
表 2钼靶、超声和远红外热成像诊断直径>2 cm乳腺病灶效能的比较(n =1193)[18]
Table2. Comparison of mammography, ultrasonography, and far-infrared thermography in the diagnosis of breast lesions large than 2 cm in diameter (n=1193)[18]
Item | Pathological examination | Accuracy/% | Sensitivity/% | Specificity/% |
---|
Malignant | Benign |
---|
Mammography | | | | | | Malignant | 354 | 116 | 85.0 | 84.9 | 85.1 | Benign | 63 | 660 | Color Doppler ultrasonography | | | | | | Malignant | 396 | 78 | 91.7 | 95.0 | 90.0 | Benign | 21 | 698 | Far-infrared thermography | | | | | | Malignant | 375 | 39 | 93.2 | 90.0 | 95.0 | Benign | 42 | 737 |
|
查看原文
表 3基于差图的乳腺癌分类结果对比
Table3. Comparison of breast cancer classification results based on differential thermal images
Item | Accuracy/% | Sensitivity/% | Specificity/% |
---|
Original images | | | | CNN | 62.8 | 23.9 | 100 | Differential thermal images | | | | CNN | 73.4 | 97.8 | 65.2 | AdaBoost | 87.2 | 91.3 | 83.3 |
|
查看原文
赵明珠, 张艳, 朱应燕. 基于红外热成像的早期疾病检测技术的研究进展[J]. 激光与光电子学进展, 2021, 58(8): 0800002. Mingzhu Zhao, Yan Zhang, Yingyan Zhu. Research Progress of Early Disease Detection Technology Based on Infrared Thermography[J]. Laser & Optoelectronics Progress, 2021, 58(8): 0800002.