Help me in raindrop removal in images. please plz......
Posts: 14,118
Threads: 61
Joined: Oct 2014
The removal of rain from a video is a difficult problem and has recently been extensively investigated. However, the problem of rain removal of a single image was rarely studied in the literature, where it is not possible to exploit temporal information between successive images, which makes the problem very difficult. This paper proposes a single - image rain - removal framework by appropriate formulation of rain removal as an image decomposition problem based on morphological component analysis (MCA). Instead of directly applying the conventional image decomposition technique, we first decompose an image into the low-frequency and high-frequency parts using a two-sided filter. The high frequency part is decomposed into "rain component" and "non-rainy component" by performing dictionary learning and poor coding. As a result, the rain component can be successfully removed from the image, retaining most of the original image details. Experimental results demonstrate the efficacy of the proposed algorithm.
In this project, we propose a single image-based rainfall elimination frame formulation for the elimination of rain stripe as an image decomposition problem based on MCA. In our method, an image is first decomposed into the low frequency and high frequency parts using a bilateral filter. The high frequency part is decomposed into "rain component" and "non-rainy component" by performing dictionary learning and sparse coding based on MCA.
The main contribution of this document is threefold:
(I) to our best knowledge, our method is one of the first to achieve the removal of rain rays while preserving geometric details in a single frame, where no temporal or movement information is required between successive images;
(Ii) we propose the first automatic frame of decomposition of images based on MCA for the removal of rain rafts; Y
(Ii) learning the dictionary to decompose rain steaks of an image is fully automatic and autonomous, where no additional training samples are required in the learning stage of the dictionary.