公开(公告)号：WO2016086034A1

公开(公告)日：2016-06-02

申请号：PCT/US2015/062500

申请日：2015-11-24

Applicant: STC.UNM

Inventor： HAYAT, Majeed, M. ,  KRISHNA, Sanjay ,  GODOY, Sebastian, Eugenio ,  RAMIREZ, David, A

IPC: A61B5/00

CPC classification number: A61B5/7275 ,  A61B5/015 ,  A61B5/489 ,  A61B5/7264 ,  G06F19/00

Abstract: A technique for classifying lesions as malignant or benign is disclosed. The technique can include: cooling an area of skin including a lesion of a patient to initiate a warm-up process; receiving a temporal sequence of thermal images of the area of skin representing a thermal recovery of the area of skin, the temporal sequence of thermal images generated by an infrared camera; generating a temporal profile of the thermal recovery based on the temporal sequence of thermal images; analyzing temporal statistical properties of the temporal profile; determining a malignancy probability that the lesion is malignant based on an analysis of the temporal profile, wherein the determining includes extracting one or more statistical features based on continuous- time stochastic signals in the sequence of thermal images; and classifying the lesion based on the malignancy probability.

Abstract translation: 公开了一种将病变分类为恶性或良性的技术。 该技术可以包括：冷却包括患者病变的皮肤区域以启动预热过程; 接收代表皮肤区域的热恢复的皮肤区域的时间序列，由红外相机生成的热图像的时间序列; 基于热图像的时间序列产生热恢复的时间分布; 分析时间分布的时间统计特性; 基于所述时间分布的分析来确定所述病变是恶性的恶性概率，其中所述确定包括基于所述热图像序列中的连续时间随机信号提取一个或多个统计特征; 并根据恶性概率对病变进行分类。