An efficient adaptive image compression using structured artificial neural network (ANN) is described. An image is first divided into a series of sub-block with size 8x8 pixels. Then each of them is transformed by a Discrete Cosine Transform (DCT) using a structured ANN. Then, all the sub-block are sorted into 4 classes using another layers of structured ANN according to their level of activity within each sub-block. Adaptivity is provided by assigning bits between classes. The neural network used is a structured one instead of a fully connected so that convergency and speed of learning are dramatically improved. Each subnetwork is trained and tested independently. Excellent performance is achieved when it compares with traditional fully connected neural network image compression methods.