基于卷積網(wǎng)絡(luò)的物體檢測(cè)應(yīng)用研究
發(fā)布時(shí)間：2018-12-21 10:59
【摘要】：提出一種基于卷積神經(jīng)網(wǎng)絡(luò)改進(jìn)的行人檢測(cè)方法。改進(jìn)主要涉及兩個(gè)方面,包括如何決定CNN樣本迭代學(xué)習(xí)次數(shù)和如何進(jìn)行重合窗口的合并。第一,關(guān)于CNN樣本迭代次序問題,在順序迭代訓(xùn)練多個(gè)CNN分類模型的基礎(chǔ)上,提出一種基于校驗(yàn)集正確率及其在迭代系列分類器中展現(xiàn)出的穩(wěn)定性并進(jìn)行更優(yōu)模型選擇的策略,以使最終選擇的分類器推廣能力更優(yōu)。第二,提出了一種不同于非極大值抑制的多個(gè)精確定位回歸框合并機(jī)制。精確定位回歸框的獲取以CNN檢測(cè)過程輸出的粗定位框作為輸入,然后對(duì)每個(gè)粗定位框應(yīng)用CNN精確定位過程并獲得對(duì)應(yīng)的精確定位回歸框,最后對(duì)多個(gè)精確定位回歸框進(jìn)行合并,合并過程考慮了每個(gè)精確定位回歸框的正確概率。更精確來說,最終的合并窗口基于多個(gè)相關(guān)的精確定位回歸框的概率加權(quán)求和方式獲得。針對(duì)提出的兩個(gè)改進(jìn),在國(guó)際上廣泛使用的行人檢測(cè)公共測(cè)試數(shù)據(jù)集ETH上進(jìn)行了一系列實(shí)驗(yàn)。實(shí)驗(yàn)結(jié)果表明,提出的兩個(gè)改進(jìn)方法均能有效地提高系統(tǒng)的檢測(cè)性能,在相同的測(cè)試條件下,融合兩個(gè)改進(jìn)的方法相比Fast R-CNN算法檢測(cè)性能提升了5.06%,達(dá)到了 40.01%的檢測(cè)結(jié)果。提出一種基于卷積神經(jīng)網(wǎng)絡(luò)的車型分類方法,首先建立了十萬數(shù)量級(jí)的不同車型不同場(chǎng)景的車型分類數(shù)據(jù)庫,通過設(shè)計(jì)網(wǎng)絡(luò)結(jié)構(gòu)并訓(xùn)練卷積神經(jīng)網(wǎng)絡(luò)模型,在其自定義測(cè)試集上進(jìn)行分類器性能驗(yàn)證。實(shí)驗(yàn)過程中對(duì)樣本進(jìn)行了對(duì)齊和擴(kuò)邊界操作,對(duì)比了不同訓(xùn)練迭代次數(shù)產(chǎn)生的CNN分類器,最終得到的CNN分類器在車型分類數(shù)據(jù)庫測(cè)試集上的平均正確率達(dá)到了 95.49%。實(shí)驗(yàn)結(jié)果表明,本文提出的卷積神經(jīng)網(wǎng)絡(luò)模型在車型分類任務(wù)上取得了較好的分類精度。
[Abstract]:An improved pedestrian detection method based on convolution neural network is proposed. The improvement mainly involves two aspects, including how to determine the number of CNN sample iterative learning and how to merge the overlap window. First, with regard to the iterative order of CNN samples, on the basis of sequential iterative training of multiple CNN classification models, a strategy based on the correct rate of check set and its stability in iterative series classifier is proposed to select a better model. In order to make the final choice of classifier promotion ability better. Secondly, a combination mechanism of multiple exact location regression frames is proposed, which is different from non-maximum suppression. The accurate location regression frame is obtained by using the coarse positioning box output from the CNN detection process as input, and then the CNN precise positioning process is applied to each coarse positioning box and the corresponding accurate positioning regression box is obtained. Finally, the multiple accurate positioning regression boxes are merged. The merging process takes into account the correct probability of each exact location regression box. More precisely, the final merge window is based on the probability weighted summation of multiple correlated precise location regression frames. Aiming at the two improvements proposed, a series of experiments have been carried out on ETH, a common test data set for pedestrian detection, which is widely used in the world. The experimental results show that the proposed two improved methods can effectively improve the detection performance of the system. Under the same test conditions, the two improved methods can improve the detection performance of the Fast R-CNN algorithm by 5.06 steps. The test results are 40.01%. This paper presents a vehicle classification method based on convolution neural network. Firstly, a model classification database of different models with different scenes of 100,000 orders of magnitude is established, and the network structure is designed and the convolutional neural network model is trained. The classifier performance is verified on its custom test set. During the experiment, the samples are aligned and expanded, and the CNN classifier produced by different training iterations is compared. The average correct rate of the CNN classifier on the test set of vehicle classification database is 95.49. The experimental results show that the proposed convolution neural network model has achieved better classification accuracy in vehicle classification task.
【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】：2017
【分類號(hào)】：TP391.41;TP183
本文編號(hào)：2388817
[Abstract]:An improved pedestrian detection method based on convolution neural network is proposed. The improvement mainly involves two aspects, including how to determine the number of CNN sample iterative learning and how to merge the overlap window. First, with regard to the iterative order of CNN samples, on the basis of sequential iterative training of multiple CNN classification models, a strategy based on the correct rate of check set and its stability in iterative series classifier is proposed to select a better model. In order to make the final choice of classifier promotion ability better. Secondly, a combination mechanism of multiple exact location regression frames is proposed, which is different from non-maximum suppression. The accurate location regression frame is obtained by using the coarse positioning box output from the CNN detection process as input, and then the CNN precise positioning process is applied to each coarse positioning box and the corresponding accurate positioning regression box is obtained. Finally, the multiple accurate positioning regression boxes are merged. The merging process takes into account the correct probability of each exact location regression box. More precisely, the final merge window is based on the probability weighted summation of multiple correlated precise location regression frames. Aiming at the two improvements proposed, a series of experiments have been carried out on ETH, a common test data set for pedestrian detection, which is widely used in the world. The experimental results show that the proposed two improved methods can effectively improve the detection performance of the system. Under the same test conditions, the two improved methods can improve the detection performance of the Fast R-CNN algorithm by 5.06 steps. The test results are 40.01%. This paper presents a vehicle classification method based on convolution neural network. Firstly, a model classification database of different models with different scenes of 100,000 orders of magnitude is established, and the network structure is designed and the convolutional neural network model is trained. The classifier performance is verified on its custom test set. During the experiment, the samples are aligned and expanded, and the CNN classifier produced by different training iterations is compared. The average correct rate of the CNN classifier on the test set of vehicle classification database is 95.49. The experimental results show that the proposed convolution neural network model has achieved better classification accuracy in vehicle classification task.
【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】：2017
【分類號(hào)】：TP391.41;TP183
【參考文獻(xiàn)】
相關(guān)期刊論文 前2條
1 馮興輝;張旭;陳禮貴;宋杰;;直線特征測(cè)量數(shù)據(jù)的平滑去噪[J];機(jī)械制造;2015年04期
2 劉遠(yuǎn)社;傅里葉變換與測(cè)不準(zhǔn)原理[J];西南民族大學(xué)學(xué)報(bào)(自然科學(xué)版);2003年05期
，本文編號(hào)：2388817
本文鏈接：http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/2388817.html
最近更新
教材專著
