AUTOMATIC SINGLE IMAGE BASED RAIN STREAKS REMOVAL VIA IMAGE DECOMPOSITION

PRESENTED BY SANDEEP T M 1KI08EC036

UNDER THE GUIDENCE OF Mr. NARENDRA T V Lecturer, ECE, Dept.

PRESENTATION OVERVIEW

Introduction. MCA based image decomposition. Rain streak removal frame work. Advantages. Application. Conclusion.

INTRODUCTION

Different weather conditions such as rain, snow, haze, or fog will cause complex visual effects of spatial or temporal domains in images or videos. Vision-based Rain Detection and Removal. Single-Image-Based Rain Streak Removal. HOG (histogram of oriented gradients).

(4 pedestrians detected)

(5

pedestrians detected)

MCA based image decomposition

Sparse coding.
Sparse coding is a technique of finding a sparse representation for a signal with a small number of nonzero or significant coefficients corresponding to the atoms in a dictionary.

Dictionary learning.
MCA algorithms usually use a fixed global dictionary based on wavelets/curvelets to represent the geometric component of an image. To represent the textural component of an image, either a fixed global (global DCT) or a local (local DCT) dictionary is used.

Rain streak removal framework

Block diagram of the rain streak removal method

LOW FREQUENCY PART

HIGH FREQUENCY PART

The input rain image decompose into HF and LF part using bilateral filters.

Detect and extract all the patches containing rain streaks from HF part.

Patches extracted from the HF part.

Rain sub-dictionary

Non-rain sub-dictionary

The intensity of each pixel in a rain streak is larger than those of the neighboring pixels outside the streak.

Rain component

Geometric component.

Based on the dictionary performing sparse coding for each patch extracted from HF part to find its sparse coefficients .

Geometric component.

LF part

Rain-removed version

By adding the non rain component and LF part result will be rain removed version.

ADVANTAGES

Image partition and de-noising. Image restoration. Preservation of geometric details in a single frame.

APPLICATION

Mobile visual search. Space object detection / recognition.

Image stitching.

CONCLUSION
A single image based rain streaks can be removed by MCA based
image decomposition method sparse coding and dictionary learning.

The dictionary learning of the proposed method is fully automatic

and sell-contained where no extra training samples are required in

the dictionary learning stage.

Extended dictionary for further enhancement of image .

REFERENCES
1.

P. C. Barnum, S. Narasimhan, and T. Kanade, Analysis of rain and snow in frequency space, Int. J. Comput. Vis., vol. 86, no. 23, pp. 256274, 2010. K. Garg and S. K. Nayar, Detection and removal of rain from videos, in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., June 2004, vol. 1, pp. 528535. K. Garg and S. K. Nayar, When does a camera see rain? in Proc. of IEEE Int. Conf. Comput. Vis., Oct. 2005, vol. 2, pp. 1067-1074. K. Garg and S. K. Nayar, Vision and rain, Int. J. Comput. Vis., vol. 75, no. 1, pp. 327, 2007. K. Garg and S. K. Nayar, Photorealistic rendering of rain streaks, ACM Trans. on Graphics, vol. 25, no. 3, pp. 996-1002, July 2006. X. Zhang, H. Li, Y. Qi, W. K. Leow, and T. K. Ng, Rain removal in video by combining temporal and chromatic properties, in Proc. IEEE Int. Conf. Multimedia Expo, Toronto, Ont. Canada, July 2006, pp. 461464.

2.

3.

4.

5.

6.

THANK YOU