本文關(guān)鍵詞： 圖文數(shù)據(jù) softmax多分類(lèi)器 多分類(lèi)支持向量機(jī) 加權(quán)KNN 融合分類(lèi)方法　出處：《中國(guó)圖象圖形學(xué)報(bào)》2016年07期 　論文類(lèi)型：期刊論文

【摘要】：目的圖文數(shù)據(jù)在不同應(yīng)用場(chǎng)景下的最佳分類(lèi)方法各不相同,而現(xiàn)有語(yǔ)義級(jí)融合算法大多適用于圖文數(shù)據(jù)分類(lèi)方法相同的情況,若將其應(yīng)用于不同分類(lèi)方法時(shí)由于分類(lèi)決策基準(zhǔn)不統(tǒng)一導(dǎo)致分類(lèi)結(jié)果不理想,大幅降低了融合分類(lèi)性能。針對(duì)這一問(wèn)題,提出基于加權(quán)KNN的融合分類(lèi)方法。方法首先,分別利用softmax多分類(lèi)器和多分類(lèi)支持向量機(jī)(SVM)實(shí)現(xiàn)圖像和文本分類(lèi),同時(shí)利用訓(xùn)練數(shù)據(jù)集各類(lèi)別分類(lèi)精確度加權(quán)后的圖像和文本正確判別實(shí)例的分類(lèi)決策值分別構(gòu)建圖像和文本KNN模型;再分別利用其對(duì)測(cè)試實(shí)例的圖像和文本分類(lèi)決策值進(jìn)行預(yù)測(cè),通過(guò)最鄰近k個(gè)實(shí)例屬于各類(lèi)別的數(shù)目確定測(cè)試實(shí)例的分類(lèi)概率,統(tǒng)一圖像和文本的分類(lèi)決策基準(zhǔn);最后利用訓(xùn)練數(shù)據(jù)集中圖像和文本分類(lèi)正確的數(shù)目確定測(cè)試實(shí)例中圖像和文本分類(lèi)概率的融合系數(shù),實(shí)現(xiàn)統(tǒng)一分類(lèi)決策基準(zhǔn)下的圖文數(shù)據(jù)融合。結(jié)果在A(yíng)ttribute Discovery數(shù)據(jù)集的圖像文本對(duì)上進(jìn)行實(shí)驗(yàn),并與基準(zhǔn)方法進(jìn)行比較,實(shí)驗(yàn)結(jié)果表明,本文融合算法的分類(lèi)精確度高于圖像和文本各自的分類(lèi)精確度,且平均分類(lèi)精確度相比基準(zhǔn)方法提高了4.45%;此外,本文算法對(duì)圖文信息的平均整合能力相比基準(zhǔn)方法提高了4.19%。結(jié)論本文算法將圖像和文本不同分類(lèi)方法的分類(lèi)決策基準(zhǔn)統(tǒng)一化,實(shí)現(xiàn)了圖文數(shù)據(jù)的有效融合,具有較強(qiáng)的信息整合能力和較好的融合分類(lèi)性能。
[Abstract]:Objective the optimal classification methods of graphic and text data in different application scenarios are different, and most of the existing semantic level fusion algorithms are suitable for the same classification methods of graphic and text data. If it is applied to different classification methods, the classification result is not ideal due to the disunity of classification decision criteria, which greatly reduces the performance of fusion classification. A fusion classification method based on weighted KNN is proposed. Firstly, image and text classification are realized by using softmax multi-classifier and multi-classification support vector machine respectively. At the same time, the KNN model of image and text are constructed by using the classification decision value of the accurate classification accuracy of each category of training data set. Then we use it to predict the decision value of image and text classification of test cases, and determine the classification probability of test cases by the number of the nearest k instances belonging to different kinds of others. Unified image and text classification decision-making benchmark; Finally, using the correct number of image and text classification in the training data set, the fusion coefficient of the probability of image and text classification in test examples is determined. The results are compared with the benchmark method and the experimental results are carried out on the image text pairs of the Attribute Discovery data set. The experimental results show that the classification accuracy of the fusion algorithm is higher than that of image and text, and the average classification accuracy is 4.45% higher than the baseline method. In addition, the average integration ability of the algorithm is 4.19% higher than that of the benchmark method. Conclusion the algorithm unifies the classification decision benchmark of different image and text classification methods. It realizes the effective fusion of graph and text data, and has strong ability of information integration and better performance of fusion and classification.
【作者單位】： 中國(guó)科學(xué)院電子學(xué)研究所;中國(guó)科學(xué)院大學(xué);
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(41301493) 高分對(duì)地觀(guān)測(cè)領(lǐng)域?qū)W術(shù)交流基金項(xiàng)目(GFEX04060103)~~
【分類(lèi)號(hào)】：TP391.41;TP18
【正文快照】： 0引言 隨著互聯(lián)網(wǎng)技術(shù)的發(fā)展,數(shù)據(jù)量呈現(xiàn)爆炸式增長(zhǎng),數(shù)據(jù)類(lèi)型不再局限于單一的文本,而是擴(kuò)展到圖像、音頻、視頻等多媒體數(shù)據(jù)。其中圖像以其豐富的視覺(jué)特征,將抽象數(shù)據(jù)直觀(guān)、生動(dòng)、形象的呈現(xiàn)給人們,使得信息的傳播和交流更為便捷。互聯(lián)網(wǎng)多媒體數(shù)據(jù)規(guī)模大、類(lèi)型多、組織結(jié)構(gòu)，

本文編號(hào)：1446001