Abstract: Early detection of breast cancer is significant for breast cancer treatment. A tumor recognition system based on tactile imaging is researched and designed, which consists of pressure signal acquisition unit, signal control unit and image processing unit. Pressure signal acquisition unit is employed to acquire pressure signal. Signal control unit sends the capacitor data of tactile sensor array received from PCap01 to host computer by EZ-USB. Image processing unit is used for filtering the noise of stress image, including least square method, Otsu method, pixel neighborhood rating based filtering and image linear interpolation. Experiment results show tumor recognition system has the ability to identify tumor in complex noisy image.

Key words: breast cancer, tactile imaging, tactile sensor array, Otsu method

摘要： 乳腺癌的早期发现对于乳腺癌的治疗有十分重要的意义，设计了基于触诊成像的肿块识别系统。该系统由压力信号检测单元、信号控制单元和图像处理单元组成，压力信号检测单元实现压力信号的采集，信号控制单元由FPGA控制多路开关的选通来扫描压力阵列传感器的电容信号，并将数据通过EZ-USB发送给上位机，图像处理单元对压力图像进行处理来滤除噪声。其中图像处理单元是该系统的枢纽，运用最小二乘曲面拟合校正压力图像，最大类间方差法和基于滤波的图像邻域评级技术对压力图像进行分割，对压力图像进行线性插值来提高图像分辨率。实验结果表明，该系统可以在复杂噪声图像中有效地识别出肿块。

关键词: 乳腺癌, 触诊成像, 触觉传感器阵列, 最大类间方差法