ICTACT Journal on Image and Video Processing

C. Rama Mohan ¹, S. Kiran ², Vasudeva ³, A. Ashok Kumar ⁴

Affiliations
1 Department of Computer Science Engineering, Visvesvaraya Technological University, IN
2 Department of Computer Science Engineering, YSR Engineering College of Yogi Vemana University, IN
3 Department of Computer Science Engineering, Shri Madhwa Vadiraja Institute of Technology and Management, IN
4 Department of Physics, YSR Engineering College of Yogi Vemana University, IN

Abstract

Multi-focus imaging fusion is a technique that puts together a fully focused object from the partly focused regions of several objects from the same scene. For producing a high quality fused image, directional selectivity and invariance characteristics are important. The ringed artifacts, however, were inserted into a fused image because of a lack of invariance and misdirection. A multi-focus image fusion algorithm is proposed to resolve these issues, in conjunction with qshiftN dual-tree complex wavelet transform and modified principal component analysis. First, the source images are translated into the FP domain. It helps in the obtaining of the row frequency components and column frequency components. Then the row-frequency elements and column-frequency elements are combined with a dual tree-complex wavelet qshiftN to transform the origin frames. Dual tree complex wavelet transforms with qshiftN has demonstrated that it provides an effective transformation for multi-resolution imaging fusion with its directional and shift-invariant characteristics. To enlarge the effectiveness of the qshiftN dual-tree complex wavelet transform in frequency partition-based method, the modified principal component analysis (MPCA) algorithm is used. The proposed fusion approach has been tested on a numeral of multi-focus images and compared to various popular methods of imaging fusion. The experimental results indicate that in subjective performance and objective assessment, the proposed fusion approach could deliver better fusion results.

Keywords

Multi-focus Image Fusion, Multi-resolution Transform, qshiftN Dual Tree Complex Wavelet Transform, Modified Principal Component Analysis, Quality Evaluation Metrics.

Full Text

     

C. Rama Mohan ¹, S. Kiran ², Vasudeva ³, A. Ashok Kumar ⁴

Affiliations
1 Department of Computer Science and Engineering, Visvesvaraya Technological University, IN
2 Department of Computer Science and Engineering, YSR Engineering College of Yogi Vemana University, IN
3 Department of Computer Science and Engineering, Shri Madhwa Vadiraja Institute of Technology and Management, IN
4 Department of Physics, YSR Engineering College of Yogi Vemana University, IN

Abstract

Multi-focus imaging fusion is a technique that puts together a fully focused object from the partly focused regions of several objects from the same scene. For producing a high quality fused image, negligible aliasing, and the ability to separate positive from negative frequencies characteristics are important. The ringed artifacts, however, were inserted into a fused image because of a lack of negligible aliasing and the ability to separate positive from negative frequencies properties. A multifocus image fusion algorithm is proposed to resolve these issues, in conjunction with multiresolution and modified principal component analysis. In this, two identical multi-focus images are considered, first they are subjected to the multi-resolution and then to the technique of modified principal component analysis. The multiresolution improves essential image features, which are best used in fusion of images, resulting in good image quality. Modified principal component analysis is applied to reduce the dimensionality of an image. The proposed fusion approach has been tested on a numeral of multifocus images and compared to various popular methods of imaging fusion. The experimental results indicate that in subjective performance and objective assessment, the proposed fusion approach could deliver better fusion results.

Keywords

Multifocus Image Fusion, Multiresolution, Modified PCA, Evolution Metrics, Image Quality.

Full Text

     

C. Rama Mohan ¹, A. Ashok Kumar ², S. Kiran ³, Vasudeva ⁴

Affiliations
1 Department of Computer Science and Engineering, Visvesvaraya Technological University, IN
2 Department of Physics, YSR Engineering College of Yogi Vemana University, IN
3 Department of Computer Science and Engineering, YSR Engineering College of Yogi Vemana University, IN
4 Department of Computer Science and Engineering, Shri Madhwa Vadiraja Institute of Technology and Management, IN

Abstract

The main issue with the multi-focus images lies in obtaining the relative information about the identification of objects in the individual images with less resolution. Hence the image fusion methods have attracted attention to obtain high resolute image with a pair of multifocus images. An attempt has been made in the present work to develop an image fusion methodology designing on multiresolution for the feature extraction and for better morphological details, the paper discussed about the Laplacian pyramid algorithm. Five sets of multifocus images obtained with different formats have been introduced to the sixteen different image fusion algorithms including the proposed method. Various statistical metrics were evaluated for each image fusion method. The careful comparison of the visual and objective metrics reveals that the proposed method shows best performance with not only having visual quality and also confirmed based on the variation of the statistical metrics.

Keywords

Multifocus Image Fusion, Multiresolution, Laplacian Pyramid, Evolution Metrics, Image Quality.

Full Text