There is OpenCV which has a large following. That may be your best starting point. You can always contribute to that in case you are looking to make a contribution.
Theory without application is meaningless. A number of times, it may become very hard to go from theory to application. If you are working on image processing, you may not have the luxury to write code that is inefficient. For example, for real-time image processing, if you are unable to analyze your frame in 33ms for a 30fps bandwidth, you might as well give up trying. The current API of OpenCV is based on C++, and it has interfaces to other languages. BTW, OpenCV was started by Intel to showcase their CPU capabilities.
Rakesh, I noticed that you are interested in image fusion. That is a neat field. You may also want to look up wavelet analysis and image registration to see how they can be used in image fusion (not just image blending). There are some nice papers by Blum and Zhang on the subject.
The openIP is an open source project of the Image Processing Group of Debrecen, University of Debrecen, to create a c++ library providing the most common methods and algorithms in the field of image processing and computer vision.
If You are interested in the library as a user or developer, feel free to access me at gyuriofkova
https://code.google.com/p/openip/
Images Java API Overview
App Engine provides the ability to manipulate image data using a dedicated Images service. The Images service can resize, rotate, flip, and crop images; it can composite multiple images into a single image; and it can convert image data between several formats. It can also enhance photographs using a predefined algorithm. The API can also provide information about an image, such as its format, width, height, and a histogram of color values.
The Camellia Library
is an open source Image Processing & Computer Vision library. Written in plain C, it is cross-platform (Unix / Linux, Windows) and robust. It already includes a lot of functions for image processing (filtering, morphological mathematics, labelling, warping, drawing, project/backproject, color conversion, loading/saving images, etc.), most of them being highly speed-optimized. It is also doxygen-documented and examples of use are provided.
Uncorrelated Multilinear Discriminant Analysis released MROGH released
SkyVoice TTS and SDK released
Uncorrelated Multilinear Principal Component Analysis released
Threshold Image for Small object released
L21 Regularized Correntropy for Robust Feature Selection released
Action Recognition by Dense Trajectories released
Sparse representation (L1 minimization) via half-quadratic minimization released
Genealized Constraints Neural Network Regression Model Subject to Linear Priors (GCNN-LP) released
Multilinear Principal Component Analysis released
Oboe: A Chinese Syntactic Parser released
Thresholding program released
Efficient Nonnegative Sparse Coding Algorithm released
Principal Component Analysis Based on Nonparametric Maximum Entropy released
Metropolis–Hastings algorithm released
CESR released
Urheen released
OpenPR-0.0.2 updated
OpenPR-0.0.2 updated
Naive Bayes EM Algorithm released
Local Binary Pattern
Code for Local Binary Pattern updated
Code for Local Binary Pattern updated
Code for OpenPR-0.0.2 updated
Code for Local Binary Pattern released
Code for Two-stage Sparse Representation released
Code for OpenPR-0.0.2 updated
Code for P3P(Perspective 3-Points) Solver released
Code for CMatrix Class released
Code for Linear Discriminant Function Classifier released
Code for NaÏve Bayes Classifier released
Code for Extended Kalman Filter released
Code for Supervised Latent Semantic Indexing released
Code for SIFT Extractor released
Code for Layer-Based Dependency Parser released
Code for HOG Feature Extraction released
Code for Quasi-Dense Matching released
Code for Probabilistic Latent Semantic Indexing released
Indeed, openCV is good library for programming in image processing. You can find in 2nd and 3nd books some open source applications: http://opencv.willowgarage.com/wiki/OpenCVBooks
Also, you can use imageJ with different plugin for initiation
CImg is definitely one of those. It is written in very clean and clear c++ code, yet it is very powerful. Everything of the library is written in a c++ header file. If you want to do some research in image processing and simultaneously enjoy c++ coding, this is an very good example:
http://cimg.sourceforge.net
another very large open source projects can be seen in the vtk and itk.
They are all pipeline implementation of the image processing and image filters.