Haider Riaz

CANDLE-J is a multi-language image denoising software designed as an ImageJ plugin and an open source alternative to the CANDLE MATLAB program. CANDLE-J is very adept at processing deep in vivo 3D multiphoton microscopy images where the signal to noise ratio is low (SNR). The denoising algorithms are written in python, Java and C. The C methods are accessed with the aide of the Java Native Access library. The Optimized Non Local Means Filter (ONLM) is the bottleneck of the CANDLE-J software and hence the ONLM algorithm is written as a multithreaded C program to improve performance. Additionally, the two C programs were compiled with the GCC optimizing option -O3 turned on. The -O3 option makes the compiler strive to improve the performance of the object code.