目的：探讨计算机图像处理技术对提高乳腺钼靶Ｘ线诊断率及对乳腺癌早诊的价值。方法：应用计算机伪彩色图像处理技术，对５０例正常乳腺和１６０例乳腺良、恶性病变典型Ｘ线图像进行处理。在此基础上总结图像特征规律，对原Ｘ线检查结果假阳性２６例和假阴性４７例Ｘ线片进行回顾性处理分析，并对１１０例患者术前ｘ线片进行前瞻性研究，对Ｔ乳腺癌和隐性乳腺癌的Ｘ线片进行处理。结果：正常乳腺和典型良、恶性病变有其独特的图像特征，可以此作为初拟诊断依据，据此标准处理后，４７例假脑性中２８例可纠正诊断，假阳性病例中，１１例腺纤维瘤中１０例，慢性炎症８例中５例，可得到纠正；１１０例前瞻性研究结果，统计处理诊断前后结果，差异有显著性。此外，图像处理可发现原Ｘ线中不宜显现的乳腺Ｔ癌和隐性乳腺癌，８例乳腺Ｔ癌，３例隐性乳腺癌处理后可确诊。结论：乳腺良、恶性病变在乳腺Ｘ线的伪彩色处理图像上有其典型特征，图像处理可提高诊断率，并在乳腺癌的早诊上有重要价值。 Objective: To investigate the value of computer image processing technology in improving the diagnostic rate of breast mammography and the value of early diagnosis of breast cancer. Methods: Using computer pseudo-color image processing technology, 50 cases of normal breast and 160 cases of breast X-ray images of benign and malignant lesions were processed. Based on this, the characteristics of images were summarized, and 26 cases of false positives and 47 false-negative X-rays were retrospectively analyzed. 110 preoperative x-ray films were prospectively studied. X-ray films for breast cancer and recessive breast cancer are processed. RESULTS: Normal breast and typical benign and malignant lesions had their unique image features. This can be used as the initial diagnosis basis. According to this standard, 28 cases of 47 cases of pseudo-cerebral rectification can be diagnosed. In false positive cases, 11 cases were diagnosed as glands. There were 10 cases of fibroids and 5 of 8 cases of chronic inflammation, which could be corrected. The results of 110 prospective studies and statistical results before and after diagnosis were statistically significant. In addition, image processing can be found in the original X-ray should not show breast T and hidden breast cancer, 8 cases of breast T-cancer, 3 cases of recessive breast cancer can be diagnosed after treatment. Conclusions: The benign and malignant breast lesions have typical features in the pseudo-color processing of breast X-ray images. Image processing can improve the diagnostic rate and have important value in the early diagnosis of breast cancer.