本論文的主旨是利用神經網路模型來辨識目標(包括辨別及分類)。我們依據各個模型的特性來處理不同的問題，並討論其優缺點及適用性。
　　論文中我們使用Sinogram和去偏極因素做為雷達目標的特徵值，利用這些特徵值來做目標辨識。
　　經由本論文實驗及數值結果，證明神經網路模型的確相當適合用於處理目標辨識之問題。

Using neural network models to recognize (including identify and classify) target is the object of the thesis. Several neural network models are applied based on their characteristics. We also discuss the performance and application of each model.
In the thesis, sinogram and depolarization factor are used as features for radar target classification in the numerical and experimental studies.
From the results given in this thesis, neural networks are demonstrated to be suitable to the problem of pattern recognition.