计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2014年
19期
178-181,191
,共5页
Tetrolet变换%脉冲耦合神经网络(PCNN)%软阈值%图像增强
Tetrolet變換%脈遲耦閤神經網絡(PCNN)%軟閾值%圖像增彊
Tetrolet변환%맥충우합신경망락(PCNN)%연역치%도상증강
Tetrolet transform%Pulse Coupled Neural Network(PCNN)%soft threshold%image enhancement
针对传统图像增强方法易损失边缘对比度以及抗噪性不强的缺点提出了一种基于Tetrolet变换与PCNN结合的图像增强方法。对待增强图像分别进行Tetrolet变换，得到不同尺度的高通和低通子带系数，并将分解后的高通子带系数进行软阈值处理；把经处理后的各尺度高通子带轮廓图像序列作为PCNN神经网络增强算子的外部输入，进而得到增强后的高通子带系数；通过Tetrolet反变换获得增强后的结果图像。数值实验结果表明，该增强算法不但能够有效抑制噪声，而且能够很好地增强图像边缘轮廓的清晰度。
針對傳統圖像增彊方法易損失邊緣對比度以及抗譟性不彊的缺點提齣瞭一種基于Tetrolet變換與PCNN結閤的圖像增彊方法。對待增彊圖像分彆進行Tetrolet變換，得到不同呎度的高通和低通子帶繫數，併將分解後的高通子帶繫數進行軟閾值處理；把經處理後的各呎度高通子帶輪廓圖像序列作為PCNN神經網絡增彊算子的外部輸入，進而得到增彊後的高通子帶繫數；通過Tetrolet反變換穫得增彊後的結果圖像。數值實驗結果錶明，該增彊算法不但能夠有效抑製譟聲，而且能夠很好地增彊圖像邊緣輪廓的清晰度。
침대전통도상증강방법역손실변연대비도이급항조성불강적결점제출료일충기우Tetrolet변환여PCNN결합적도상증강방법。대대증강도상분별진행Tetrolet변환，득도불동척도적고통화저통자대계수，병장분해후적고통자대계수진행연역치처리；파경처리후적각척도고통자대륜곽도상서렬작위PCNN신경망락증강산자적외부수입，진이득도증강후적고통자대계수；통과Tetrolet반변환획득증강후적결과도상。수치실험결과표명，해증강산법불단능구유효억제조성，이차능구흔호지증강도상변연륜곽적청석도。
In order to solve the problem that the traditional image enhancement method is easy to damage the edge contrast and the shortcoming of noise resistance is not strong, this paper proposes an image enhancement method based on Tetrolet transform combined with PCNN. The different scales of high pass and low pass sub-bands are obtained by Tetrolet trans-form. Then the soft threshold processing of the decomposition image high-pass sub-bands coefficients is conducted. Finally, the reconstruct image is obtained by the inverse Tetrolet transformation. The experimental results show that this enhance-ment algorithm not only can suppress noise effectively, but also can enhance the sharpness of the image edge contour.
