From rivas.ed at gmail.com Fri Oct 7 13:06:53 2005 From: rivas.ed at gmail.com (Edward Rivas) Date: Fri, 7 Oct 2005 12:06:53 -0500 Subject: [Physionet-users] problems installing WFDB software Message-ID: Let me cut to the chase... -I've installed cygwin succesfully(at least I think I did...) and included libcurl. -I've installed xview and made the suggested changes to the path and such -I downloaded wfdb 10.3.17 and unpacked the wfdb.tar.gz file into /home/wfdb -I cd'ed to /home/wfdb and entered ./configure into the command line --I selected d for a dynamically loaded library. there is nothing in the web instructions about this. how does one make a selection? --next it asked where to install WFDB... I had already tried entering my home directory and it didn't work so I just hit enter for the default, /usr. again, nothing in the web instructions mentions this aspect. --same for MANDIR, default to /usr/man --it says it did find libcurl and XView libraries. --it says that wfdb-10.3.17-cygwin is now ready to be compiled using gcc, some other stuff and then "WFDB software configuration is complete" --the command line instructs me to now run "make" then "make install" then "make check". the website instructions say to run "make install", skipping mention of "make". this conflicts with the command line. I'm not a unix user so whether this is an important distinction or which is the correct way to do it is not obvious to me. -I ran "make install" --there were errors at the end -I ran "make check" --more errors here are the last few lines of the "make install" output, up to this point everything else looks okay: /usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld: cannot find -lssl collect2: ld returned 1 exit status make[1]: *** [all] Error 1 make[1]: Leaving directory `/home/wfdb/wfdb-10.3.17/lib' make: *** [install] Error 2 can anyone make sense of this and provide a direction in which to go to fix it? Apologies for the long post, I just wanted to make that what I'd done to this point was clear. regards, Edward From rasch at med1.med.tum.de Fri Oct 7 15:40:41 2005 From: rasch at med1.med.tum.de (Raphael Schneider) Date: Fri, 7 Oct 2005 21:40:41 +0200 Subject: [Physionet-users] problems installing WFDB software In-Reply-To: References: Message-ID: <200510072140.41526.rasch@med1.med.tum.de> On Friday 07 October 2005 19:06, Edward Rivas wrote: > /usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld: cannot > find -lssl > collect2: ld returned 1 exit status > make[1]: *** [all] Error 1 > make[1]: Leaving directory `/home/wfdb/wfdb-10.3.17/lib' > make: *** [install] Error 2 I have nearly no experience with cygwin but it seems that libssl is missing. I checked the website of cygwin and found 'openssl' and 'openssl-devel' packages. Are these packages installed on your system? Raphael From naresh.bajaj at gmail.com Wed Oct 26 11:04:47 2005 From: naresh.bajaj at gmail.com (naresh bajaj) Date: Wed, 26 Oct 2005 10:04:47 -0500 Subject: [Physionet-users] MIT malignant database Message-ID: <7ff6fe720510260804j42c9c6e7q3c6666c9e61ad0df@mail.gmail.com> Hello all, I am working on MIT-SCD and MIT-malignant database. As far as I know, MIT-malignant database is obtained from MIT-SCD holter recordings. I am trying to extract all the VT and VF episodes from the database. I want beat by beat annotation as well as rhythm annotations to test my algorithm. I believe MIT-malignant database have rhythm annotation but not beat by beat annotations, while MIT-SCD has beat by beat annotation but not VT,VF rhythm annotation. As MIT-malignant database are the part of MIT-SCD database, I would like to know whether we have any information on file name and time information that were considered to create MIT-malignant recordings, so that I can map back to MIT-SCD database and get the rhythm as well beat by beat annotation. Looking forward to your reply, Thanking you, Naresh -- Naresh Bajaj, University of Calgary, 403-891-7150 From george at mit.edu Wed Oct 26 17:55:40 2005 From: george at mit.edu (George Moody) Date: Wed, 26 Oct 2005 17:55:40 -0400 Subject: [Physionet-users] MIT malignant database In-Reply-To: <7ff6fe720510260804j42c9c6e7q3c6666c9e61ad0df@mail.gmail.com> References: <7ff6fe720510260804j42c9c6e7q3c6666c9e61ad0df@mail.gmail.com> Message-ID: <435FFB5C.3080307@mit.edu> naresh bajaj wrote: [...] > As MIT-malignant database are the part of MIT-SCD database, I would like to > know whether we have any information on file name and time information that > were considered to create MIT-malignant recordings, so that I can map back > to MIT-SCD database and get the rhythm as well beat by beat annotation. Sorry, this information was not recorded, but it should be recoverable with a modest amount of effort if you follow the suggestions below. To clarify the question for other readers of this discussion: The two databases are the MIT-BIH Malignant Ventricular Arrhythmia Database, at http://physionet.org/physiobank/database/vfdb/ which contains half-hour excerpts from the records of the Sudden Cardiac Death Holter Database, at http://physionet.org/physiobank/database/sddb/ . The MIT-BIH Malignant Ventricular Arrhythmia Database includes rhythm annotations, but no beat annotations. The Sudden Cardiac Death Holter Database includes unaudited (computer-generated) beat annotations for all 23 records, and audited (manually-corrected) beat annotations for 12 of these records; these do not include rhythm annotations. The original analog tape recordings from which both databases were derived were digitized twice (once in the 1980s and again about 10-15 years later). It is not trivial to match the rhythm annotations from the older database with the beat annotations from the other, because of tape speed variations. Nevertheless, it should not be too hard. A reasonable approach might be to use a QRS detector such as 'wqrs' (available in the WFDB software package) to create beat annotations for both sets of records, and then to plot RR interval tachograms from these annotations on paper at the same time scales. It should then be fairly straightforward to match the record pairs visually, based on the tachograms. To see how to make tachograms, read PhysioNet's HRV tutorial at http://physionet.org/tutorials/hrv/ (Note that it's not necessary to correct any annotation errors that the QRS detector makes if the goal is simply to recover the information needed to pair up and time-align the two sets of records, since one would expect that these errors would also match reasonably well if generated from two versions of the same original analog tape. Of course, correcting the errors would become important if one planned to use these annotations for any other purpose.) If you try this, please share your findings with us! Because of the complexity of these recordings, annotating them is difficult and tedious. Given their significance and scarcity, any progress on this ongoing task is likely to be of considerable interest and benefit to the community. We welcome contributions of annotations as well as additional recordings to the Sudden Cardiac Death Database. -- George From h_gholizade at yahoo.com Thu Oct 27 03:53:35 2005 From: h_gholizade at yahoo.com (hossein gholizade) Date: Thu, 27 Oct 2005 00:53:35 -0700 (PDT) Subject: [Physionet-users] Reentry Message-ID: <20051027075335.87972.qmail@web31004.mail.mud.yahoo.com> Hi all, I'm going to work on reentry. let me know is there any one that has worked on it. Thanks --------------------------------- Yahoo! FareChase - Search multiple travel sites in one click. From pmijangos at pampano.unacar.mx Mon Oct 31 16:52:02 2005 From: pmijangos at pampano.unacar.mx (pmijangos at pampano.unacar.mx) Date: Mon, 31 Oct 2005 15:52:02 -0600 (CST) Subject: [Physionet-users] question about rr2ann command Message-ID: <33133.10.3.0.214.1130795522.squirrel@delfin.unacar.mx> Hello, I hope you have some hint for the next trouble >From WFDB applications guide, at the end of the description of the command ?wrsamp? it is exemplified the following rdsamp -r 100s | wrsamp -o 100w -F 360 1 2 ? This command creates a record named ?100w? that is a copy of record ?100s? (although the signal file format is different).? However, at the time of creating a header file 100w.atr using the command ?rr2ann? it is not possible for me to get a right plot of the data using the command ?hrplot?. In other words, it seems that the graphical properties of the copy are not the same than that of the original 100s file. I get a decreasing function. On the other hand, the ?rdsamp? command works fine, reading exactly the same from 100s or 100w files. I have installed the binary version of the WFDB package in Redhat-Fedora 3. Best regards, Pavel Ritto Universidad Aut?noma del Carmen Cd. del Carmen, Campeche M?xico