C / C++ / MFC

16-Jan-10 16:18

Hey all,

I want to implement a program in an embedded system. I want to write a program which simulates a human heart for the purposes of detecting a heart attack. The model of the heart should be running, then a second program will analyze this information output from the model, and when conditions are met, indicate that a heart attack is occurring.

How should I approach the modeling of the heart. Maybe EKG waves is a good indicator of a heart attack, however, what the best way to model EKG activity in an embedded system?

Thank you.

Tim Craig16-Jan-10 21:02

16-Jan-10 21:02

I smell a homework problem here although an interesting one. If you're using an embedded system to actually detect a heart attack, what do you think you'd use as input? What has your literature search turned up about EKG and heart attack?

You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.

user_void17-Jan-10 5:55

17-Jan-10 5:55

So far the research I have done and books on EKG I am currently reading indicates that impending myocardial infarction (a heart attack) is related to a rise in the ST wave segment of the EKG (ST elevation). SO it seems simple enough to model only the ST segment of the EKG. Maybe using a random number generator? Suggestions are welcome.

Bram van Kampen17-Jan-10 6:09

Bram van Kampen

17-Jan-10 6:09

user_void wrote:
A Hart Attac is related to a rise in the ST wave segment of the EKG (ST elevation). SO it seems simple enough to model only the ST segment of the EKG.

Hi

You are asking a question on a CPP/MFC forum. Not all of us have a degree in medicine. What does the above mean in plain English.

I assume that your data arrives in a continuous bite stream, which is being buffered somewhere, and that your code will read the buffer, and do it's thing.

My gut feeling is that your first step should be to do some form of Fourier analysis on the Data. Details on how to do that can be found by googeling Digital Data Processing. A word of warning: this is NOT a trivial subject!

Regards, Smile | :)

Bram van Kampen

Tim Craig17-Jan-10 14:04

17-Jan-10 14:04

It would seem that a random data stream would not be a sufficient model of anything other than a random stream of data. An instrument looking for a heart attack has to know both what is normal and what is abnormal. If it's normal, don't mess with it and if it isn't, sound an alarm or give the heart a nice jolt and get it started again. Of course, with this problem statement, not evertthing abnormal is a heart attack. And I know from experience that not all heart attacks involve the searing pain down your left arm. Sigh | :sigh:

You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.

user_void17-Jan-10 14:23

17-Jan-10 14:23

Okay, lets assume we are modeling the ST-segment elevations. The program knows what values are abnormal and will only sound an alarm when an abnormal condition is detected. If the range of this number is from 0-10, and anything greater than 4 is a sign of myocardial infraction. And if a rnd number generator is not the best solution for modeling the ST-segment elevation, then what would be the best solution?

Tim Craig17-Jan-10 17:21

17-Jan-10 17:21

To step back, is a heart attack the only reason the number climbs above a 4? I would expect EKGs are graphs for a reason and it's a waveform that suggests the heart attack. But I'm not an expert on EKGs, I've seen a number so I know what they look like and what they're used for but I haven't studied their signal signatures in detail. I just suspect that it's more than a simple threshold problem.

If the instrument is a simple monitor, a occasional false positive isn't a huge problem if they're reasonably infrequent. However, if this is part of a defribrillator that's meant to be used by medically untrained personnel or to be implanted and operate automatically, shocking a normally beating heart probably isn't the greatest idea in the world.

You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.

user_void18-Jan-10 8:46

18-Jan-10 8:46

Yes, usually when a specific segment of the EKG (ST-segment) rises above a threshold number, it indicates impending myocardial infraction. Perhaps I had structured the question incorrectly to begin with. Since I am doing all software, I an not designing an EKG machine, nor am I using an existing EKG machine for an input interface. What I want to model is the output of an EKG machine. Not even the entire output waveform of an EKG, but a specific segment (as mentioned above). I'm just trying to make the modeling as simple as possible. Thus far, I can only think of random number generator.

Tim Craig18-Jan-10 9:16

18-Jan-10 9:16

If all it takes is looking at a stream of data in the range of 0-10 and noticing when a value rises above 4, the problem is trivial. A random number generator would work just fine. Or if you wanted to get a little fancier, try to model the waveform and throw in some random "noise" which occasionally rises to the level of a heart attack.

You measure democracy by the freedom it gives its dissidents, not the freedom it gives its assimilated conformists.

header file

khomeyni16-Jan-10 9:15

khomeyni

16-Jan-10 9:15

hello everybody
when i include a header file with space between words of its name then my compiler cannot link them and take error:
[sajad@sajad binary search trees]$ g++ binary\ trees.cpp
binary trees.cpp:3:73: error: /home/sajad/src/ds/test/binary\ search\ trees/BST_node_base.h: No such file or directory
binary trees.cpp:4:63: error: /home/sajad/src/ds/test/binary\ search\ trees/BST.h: No such file or direct

while the directory and address is completely true.

#include<iostream>
#include "/home/sajad/src/ds/test/binary\ search\ trees/BST_node_base.h"
#include "/home/sajad/src/ds/test/binary\ search\ trees/BST.h"

but when i change the name ,without space between it then it run correct,what can i do to solve this error?

also when i include a fileheader and in it use for example cout<<"hello"<<endl; takes="" error,but="" when="" i="" include<iostream=""> using namespace std it work true.does it must be as i write or there is another way to solve this?

this is the part of header file:

#include<iostream>
using namespace std;
template <class T>class BS_Tree{
public:
bool direction;
int n;
BST_node<T>* root;
BS_Tree(){
root=NULL;
}
....

image processing toolkits | batch image processing

C^hris L^osinger16-Jan-10 9:47

C^hris L^osinger

16-Jan-10 9:47

khomeyni wrote:
what can i do to solve this error?

what error ?

why would you put spaces between the parts of the path ?

CPallini16-Jan-10 10:26

CPallini

16-Jan-10 10:26

You may either avoid spaces inside the paths or use relative ones.
As about the namespace,

You have two options:

specify, as you did, using namespace std;
Qualify each (std) object with the std namespace, e.g. std::cout instead of cout, std::endl instead of endl etc...

Stuart Dootson16-Jan-10 11:10

Stuart Dootson

16-Jan-10 11:10

Take out the back-slashes that quote the spaces. That is, change this:

#include "/home/sajad/src/ds/test/binary\ search\ trees/BST_node_base.h"
#include "/home/sajad/src/ds/test/binary\ search\ trees/BST.h"

to this

#include "/home/sajad/src/ds/test/binary search trees/BST_node_base.h"
#include "/home/sajad/src/ds/test/binary search trees/BST.h"

Oh - but don't, don't, don't use absolute paths in code like that! What happens when the code-base moves? You're screwed, aren't you.

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

MVP for 2010 - who'd'a thunk it!