本文關(guān)鍵詞：基于病理圖像的乳腺腫瘤定量化分析　出處：《南京信息工程大學(xué)》2016年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 乳腺癌 HE染色病理圖像 細(xì)胞檢測與分割 病理分級(jí) 預(yù)后分析

【摘要】：在乳腺癌診斷與預(yù)后過程中,通常由醫(yī)生通過顯微鏡觀察組織切片中不同的病理標(biāo)志物對(duì)病理等級(jí)進(jìn)行評(píng)分。然而人工分析的方式耗時(shí)且?guī)в休^強(qiáng)的醫(yī)生主觀性,不同醫(yī)生的診斷結(jié)果存在不一致性,這可能會(huì)給患者帶來嚴(yán)重的“過度治療”和“治療不當(dāng)”。因此,研究計(jì)算機(jī)輔助診斷系統(tǒng)能為臨床醫(yī)生提供準(zhǔn)確、定量的輔助分析結(jié)果從而加快治療進(jìn)程,對(duì)于醫(yī)生和患者都具有重要意義。諾丁漢分級(jí)系統(tǒng)與細(xì)胞的外觀和空間分布特征存在密切的聯(lián)系,因此病理圖像中細(xì)胞的檢測與分割是構(gòu)建病理圖像自動(dòng)分析系統(tǒng)的基礎(chǔ)。然而,病理圖像組織結(jié)構(gòu)復(fù)雜成分眾多,且細(xì)胞外觀呈現(xiàn)高度的異質(zhì)性。此外細(xì)胞之間存在重疊、擠壓現(xiàn)象,因此細(xì)胞的檢測與分割是一項(xiàng)極具挑戰(zhàn)性的工作。針對(duì)這一問題,本文提出了基于深度卷積神經(jīng)網(wǎng)絡(luò)初始化的主動(dòng)輪廓自適應(yīng)橢圓擬合細(xì)胞分割方法。該方法運(yùn)用卷積神經(jīng)網(wǎng)絡(luò)結(jié)合滑動(dòng)窗口自動(dòng)檢測細(xì)胞,根據(jù)細(xì)胞檢測結(jié)果初始化主動(dòng)輪廓模型,最后使用自適應(yīng)橢圓擬合方法分割重疊的細(xì)胞。為了驗(yàn)證該方法的性能,本文分別在三個(gè)數(shù)據(jù)集上進(jìn)行了測試。實(shí)驗(yàn)結(jié)果表明：本文方法在三個(gè)數(shù)據(jù)集上的檢測準(zhǔn)確率分別為：73.33%,83.91%和76.88%,在數(shù)據(jù)集1和2上分割準(zhǔn)確率分別為：85.03%,90.33%,說明本文方法性能優(yōu)于其他對(duì)比方法。在自動(dòng)病理分級(jí)研究中,本文提出了一個(gè)基于多特征描述的乳腺腫瘤病理自動(dòng)分級(jí)方法。該方法使用卷積神經(jīng)網(wǎng)絡(luò)模型檢測病理圖像中的上皮細(xì)胞和淋巴細(xì)胞；然后運(yùn)用顏色分離算法把細(xì)胞通道從HE染色的病理圖像中分離,接下來使用自適應(yīng)閾值、形態(tài)學(xué)操作、帶有前景標(biāo)記的分水嶺算法和橢圓擬合得到細(xì)胞的邊界。隨后提取出細(xì)胞的形狀紋理和反映細(xì)胞分布的空間結(jié)構(gòu)特征,將這些特征降維后輸入到支持向量機(jī)中實(shí)現(xiàn)對(duì)病理圖像自動(dòng)分級(jí)。實(shí)驗(yàn)結(jié)果表明：本文方法整體分類準(zhǔn)確率為90.20%,對(duì)高中低各等級(jí)的區(qū)分準(zhǔn)確率分別為92.87%,82.88%和93.61%,其性能遠(yuǎn)高于其他對(duì)比方法。
[Abstract]:In the diagnosis and prognosis of breast cancer. The pathological grade is usually scored by the doctor by observing the different pathological markers in the tissue slice by microscope. However, the manual analysis method is time-consuming and has a strong subjectivity of the doctor. The results of different doctors' diagnosis are inconsistent, which may bring serious "overtreatment" and "improper treatment" to patients. Therefore, the study of computer-aided diagnosis system can provide clinicians with accuracy. Quantitative analysis of the results to speed up the treatment process is of great significance to both doctors and patients. Nottingham grading system is closely related to the appearance and spatial distribution of cells. Therefore, the detection and segmentation of cells in pathological images is the basis of constructing an automatic analysis system for pathological images. However, there are many complex components in pathological images. Moreover, the appearance of cells is highly heterogeneous. In addition, there is overlap and squeezing between cells, so the detection and segmentation of cells is a very challenging task. In this paper, an active contour adaptive ellipse fitting cell segmentation method based on deep convolution neural network initialization is proposed, which uses convolution neural network combined with sliding window to automatically detect cells. The active contour model was initialized according to the results of cell detection, and then the overlapping cells were segmented by adaptive ellipse fitting method to verify the performance of the method. The experimental results show that the accuracy of this method on the three datasets is 83.91% and 76.88%, respectively. The segmentation accuracy on data set 1 and 2 is: 85.03 / 90.33, respectively, which shows that the performance of this method is superior to that of other comparison methods. In this paper, an automatic classification method for breast tumors based on multi-feature description is proposed, which uses convolution neural network model to detect epithelial cells and lymphocytes in pathological images. Then the color separation algorithm is used to separate the cell channel from the pathological image stained by HE. Then the adaptive threshold and morphological operation are used. The boundary of cells was obtained by watershed algorithm with foreground marker and ellipse fitting. Then the shape and texture of cells and the spatial structure characteristics reflecting the distribution of cells were extracted. These features are reduced and then input into support vector machine to realize the automatic classification of pathological images. The experimental results show that the overall classification accuracy of this method is 90.20%. The accuracy of distinguishing between high and low grades is 92.87% and 93.61%, respectively, and its performance is much higher than that of other comparison methods.
【學(xué)位授予單位】：南京信息工程大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41;R737.9

【相似文獻(xiàn)】

相關(guān)期刊論文 前8條

1 何建芳,劉旭明;應(yīng)用圖像處理技術(shù)提高病理圖像質(zhì)量[J];診斷病理學(xué)雜志;2002年01期

2 王偉智,劉秉瀚,鄭智勇;基于形態(tài)特征的淋巴瘤病理圖像分級(jí)分割方法[J];中國體視學(xué)與圖像分析;2005年01期

3 宋來鳳,彭勃,張義華,王亞光,梁鳳玲;心血管病理圖像的微機(jī)計(jì)量[J];中國循環(huán)雜志;1992年01期

4 劉秉瀚,王偉智,鄭智勇,黃慶榮;病理圖像中重疊細(xì)胞自動(dòng)分離的研究[J];中國體視學(xué)與圖像分析;2002年01期

5 劉秉瀚,鄭智勇,王偉智;肝組織免疫組化病理圖像的自動(dòng)分析[J];中國體視學(xué)與圖像分析;2000年04期

6 李志宏,史元春,周智明;Web上基于內(nèi)容的病理圖像檢索的研究方法與展望[J];第四軍醫(yī)大學(xué)學(xué)報(bào);1999年03期

7 徐蕓,文小崗,王振豫,李繼昌;脂肪肝CT圖像、肝穿活檢病理圖像定量形態(tài)測量軟件的設(shè)計(jì)和應(yīng)用[J];中華放射學(xué)雜志;2005年10期

8 ;[J];;年期

相關(guān)會(huì)議論文 前3條

1 李志宏;史元春;;Web上基于內(nèi)容的病理圖像檢索的設(shè)計(jì)與展望[A];信息科學(xué)與微電子技術(shù):中國科協(xié)第三屆青年學(xué)術(shù)年會(huì)論文集[C];1998年

2 鄭唯強(qiáng);;虛擬顯微鏡技術(shù)在病理學(xué)的應(yīng)用[A];中華醫(yī)學(xué)會(huì)病理學(xué)分會(huì)2006年學(xué)術(shù)年會(huì)論文匯編[C];2006年

3 鄭清平;周澤斌;;一款較適用于病理圖像數(shù)字化處理的數(shù)碼相機(jī)[A];面向二十一世紀(jì)的生物醫(yī)學(xué)體視學(xué)和軍事病理學(xué)論文摘要匯編[C];2000年

相關(guān)碩士學(xué)位論文 前6條

1 王冠皓;深度卷積網(wǎng)絡(luò)及其在乳腺病理圖像分析中的應(yīng)用[D];南京信息工程大學(xué);2015年

2 龔磊;基于病理圖像的乳腺腫瘤定量化分析[D];南京信息工程大學(xué);2016年

3 趙明珠;細(xì)胞病理圖像的特征分析與分類識(shí)別[D];浙江工業(yè)大學(xué);2012年

4 張建波;基于流行學(xué)習(xí)的淋巴瘤組織病理圖像分類研究[D];福州大學(xué);2011年

5 張麗熙;基于紋理特征空間的淋巴組織病理圖像協(xié)同分類研究[D];福州大學(xué);2006年

6 王玉山;基于神經(jīng)網(wǎng)絡(luò)的病理圖像融合識(shí)別研究與實(shí)現(xiàn)[D];武漢理工大學(xué);2010年

，

本文編號(hào)：1412402