摘要
讨论神经网络在图象压缩编码领域中的应用.针对图象矢量量化存在的分块效应问题,对Kohonen自适应模型进行了研究.应用了两个DCT的余弦变换域的特征值,结合Kohonen的自组织特征映射(SOFM)算法对图象元素分类压缩编码.计算机模拟实验表明,和单纯用神经网络直接进行矢量量化相比,应用这种技术的图象编码压缩比和译码图象质量都有明显的提高.
The application of neural network in image compression coding is discussed. Based on blocking effect of vector quantization of image,Kohonen's self-organizing feature maps(SOFM) was stud-ied.Image element was classified compression coding with two eigenvalues of DCT (Discrete Cosine Trans-form) and to combine SOFM algorithm. The results of computer simulation show compression coding of im-age quality comparing with. direct vq and compression coding of image are grently improved.
出处
《湖北师范学院学报（自然科学版）》
1997年第6期6-8,26,共4页
Journal of Hubei Normal University(Natural Science)
