ABSTRACT
This paper proposes a scheme for noise reduction in low illumination images. There are various methods already in existence for noise reduction but hybrid models have given better results many times. The proposed scheme is based on a Hybrid Neuro-Fuzzy Model. This model may be termed as Fuzzy Convolutional Neural Network. While detecting and recognizing texts present in the scene images, there is an issue of noise. This noise may be present due to low intensity of light. Since the proposed model will be an integration of Neural Network and Fuzzy Logic, hence it will have the capability to learn and to handle uncertainties present in the noise affected image. This model will reduce the noise present in the image and will present a better recognition of texts in scene images.
