摘要：用科學(xué)預(yù)測模型對乳房運(yùn)動(dòng)軌跡進(jìn)行準(zhǔn)確的預(yù)測,可以節(jié)約大量的計(jì)算和分析時(shí)間。分別建立BP和GRNN神經(jīng)網(wǎng)絡(luò)模型,將乳頭點(diǎn)的運(yùn)動(dòng)坐標(biāo)作為輸入值,預(yù)測乳房其他4個(gè)部位的運(yùn)送軌跡。結(jié)果表明,兩種預(yù)測模型都可以較好地實(shí)現(xiàn)乳房運(yùn)動(dòng)軌跡的預(yù)測。BP神經(jīng)網(wǎng)絡(luò)可以預(yù)測出乳房4個(gè)測量點(diǎn)的運(yùn)動(dòng)軌跡,但要求樣本量必須足夠大,且網(wǎng)絡(luò)參數(shù)不好確定,乳房軌跡預(yù)測時(shí)所需時(shí)間較長,效率低;GRNN神經(jīng)網(wǎng)絡(luò)克服了BP神經(jīng)網(wǎng)絡(luò)的缺點(diǎn),預(yù)測值更接近真實(shí)值,4個(gè)測量點(diǎn)預(yù)測值分別比BP神經(jīng)網(wǎng)絡(luò)高出5.95%,5.33%,6.37%,6.97%,可以較好地預(yù)測乳房運(yùn)動(dòng)軌跡,其預(yù)測值分別達(dá)到了真實(shí)均值的94.68%,93.87%,93.76%,94.79%。