A novel image de-noising algorithm combined ICM with morphology

After summarizing the property of pulse noise, a de-noising algorithm combined Intersecting Cortical Model (ICM), a simplification of Pulse-Coupled Neural Network (PCNN) model, with morphology is proposed in this paper. Firstly noisy points are located using the synchronous pulse burst property of ICM, then, processed using the morphological filtering. Finally our algorithm is compared with median filtering, morphological filtering and the algorithm combined PCNN with median filtering. The ICM is combined with mathematical morphology. The experimental results show that our algorithm performs well and is able to catch up with or even exceed the algorithm combined PCNN with median filtering. Also, ICM is faster than PCNN and hence our algorithm is preferable when possible.