Arrhythmia Detection and Classification C++ implementation programme and algorithm accuracy improvement

Dear Researcher / Scientist, We have started to prepare ECG simulator that display MITDB records from [login to view URL] ECG files and detect as much ( PCV, etc. ) arrhythmia types in it. You goal will be to modify existing C++/Qt project or to write new one that will detect and classify all known arrhythmia types in input ECG waves. ( You may refer to works of [login to view URL] as well ) I've investigated part of your R&D works on the domain of ECG data analysis presented at [login to view URL] and at codeproject and have found them singnificant and utilizable. I have a project to complete, to code an application for Windows firstly, than from Linux and more platforms including embedded and mobile. The application should open the physionet ECG data files, preprocess them, extract next parameters in real time ( on a particular specified time frame - window ) and offline: calculate the P, Q, R, S, T, U peak and P-R interval, P-R segment, S-T segment, S-T interval, and QRS interval. I have already completed about 45 % of the project, extracted QRS intervals, RR intervals, R peaks, Hear Rate ( BPM ) using unique ( proprietary ) algorithm. The application is coded in C++ with usage of Qt toolkit, that makes it extremely portable and utilizable. I have a client for the application from New York and I will sell it after completion for a significant budget. Please see the work completed till now is in archive: [login to view URL] What is left to be done is to compute Q,P,S,T,U peaks, than PR interval and segments with higher accuracy, ST intervals and segments ( actually all the parameters for the systole ). Than what is also needed is to code robust arrhythmia detection and classification module:Based on the calculated parameters, classify arrhythmia. (For this we can use SVM or NN. But for this, we need large amount of ECG and arrhythmia data to training the algorithm. But we may use MIT ECG DB.)Otherwise, instead of machine learning technique, if there is another way to classify arrhythmia, then we can try.I have found already lots of data mining and machine learning methods to detect and classify arrhythmia, but most of them are quite complex and not [login to view URL] need as simple and comprehensive C++ implemented algorithm that will receive an array of samples ( double * ) and extract the above parameters + conclude on arrhythmia class ( classify the rhythm ) presented within the samples. That is all.I know you are quite experiences on this domain and can prepare this algorithm. If you are interested to code it please let me know can you make such C++ function that will accept array of double - the samples and will return rhythm type and the above parameters. For more clarifications, information do not hesitate to contact me at gmail / google talk :Thanks in advance for help within the undertaking, Target platform could be Windows or Ubuntu. You may can easily evaluate the code. If the parameters for MITDB and simulator data are different, we may need to separate the parameters or separate code. Please check. The file name is the arrythmia name and 3-10-3 is the 3 sec normal 10 sec arrythmia 3 sec normal. *** if we can display the annotation on each case, it would be great and easily compare the performance. If the simulated data is too short, please let me know. I will generate more. The code compare each beat by beat and see if the beat is arrythmia or not. I think we can add one more alarm sign, This sign would be on when the arrythmia beat occurs and off when the normal beat occurs. Please let me know your idea. if you are looking for the MIT data, then link is as below *** The project is about to build ECG monitoring simulator. We have many ECG data. The simulator load one of these data and show as the real time ECG signal. If some event occurs, then load the other ECG data. prefer c++ or c#. After confirmed and ability to complete is proved I can upload the existing C++ source codes to you. +NDA