目的:新疆地区女性多为脂肪型乳腺,B超回声影像具有很强的特殊性,当前国内外的深度学习模型识别率偏低。方法:为提升新疆地区乳腺超声影像的识别精确度,提升辅助诊断效果,本文提出了一种基于U-Net网络的超声乳腺肿瘤的分割方法UBDR-FB(ultrasound big data recognition for fatty breast)。UBDR-FB包含编码和解码两部分,在编码的过程中,UBDR-F
Modern book design has various forms.In order to give readers a better reading experience,the use of comprehensive materials is a common element in books,which can not only enrich the sense of touch a