基于脉冲耦合神经网络和施密特正交基的一种新型图像压缩编码算法

自从脉冲耦合神经网络(PCNN)被提出以来,在图像处理、模式识别、人工智能等领域得到了广泛应用.由于其生物学背景的特性,使得其能够对灰度图像进行完美的分割:PCNN局部连接域的作用及阈值指数衰减特性,使得具有近似灰度特性的邻近像素能够同时处于激活状态,这就构成了PCNN分割特性的基础,使得图像分割结果既能较好地包含原始图像细节信息,又能避免一些无意义的小分割块的产生.借鉴施密特正交化思想,利用自然初始基对每一分割区域进行变换,得到一组正交基的变换系数,相对于分割前图像的数据量大为减少,存储空间需求小,从而实现了压缩.相对于JPEG算法,该方法使重建图像的质量得到显著提高,同时也使得逐步重建图像成为可能.

A Novel Image Compression Coding Algorithm Based on Pulse-Coupled Neural Network and Gram-Schmidt Orthogonal Base

Pulse Coupled Neural Network(PCNN) has gained widely application in image processing,pattern recognition,artificial intelligence etc,since it was proposed.PCNN can perform perfect image segmentation due to its biological background.PCNN has the property of local interconnection and changing threshold through which those adjacent pixels that have approximate gray values can be pulsed simultaneously.So PCNN has the foundation of realizing the regional segmentation.And segmented images that contain the details of origin can be achieved and at the same time the trivial segmented regions may be avoided.For the better approximation of irregular segmented regions,the Gram-Schmidt method,by which a group of orthogonal base functions is constructed from a group of linear independent initial functions,is adopted.It was found that much less computer memory was needed to store the coefficients than the original image,resulting in data compression.Because of the orthogonal reconstructing method,the quality of reconstructed image can be greatly improved and the progressive image transmission also becomes possible compared to JPEG algorithm.