Distributed Image compression in wireless sensor networks using intelligent methods

ABSTRACT: The availability of low cost hardware like microphone and CMOS cameras has led to spread in multimedia wireless sensor network. Data transfer is considered one of the high cost tasks in wireless sensor network consuming more than 80 percent of dedicated energy. The most appropriate way to reduce energy consumption is to use an appropriate compression image technique. Designing a compression method requires balance and careful estimation among compression ratio factors, quality, complexity, and time. The present article presents a new method, unlike previous ones, focusing only on the structure of the network. In addition, it tries to deal with low compression rate along with high quality of recovered images and reduction in energy consumption. Hence, the method used here is Distributed Vector Quantization (DVQ) whose codebook, being real-time and dynamic, was created by applying neural network of adaptive resonance theory. The results from tests on primary images demonstrate the accuracy and efficacy of proposed method.