D120.de/forum

Hi, I have a question regarding task 1.f: We should compute the 20-step ahead prediction. The target state is 0 (I thought this means q1_t=0 and q2_t=0.) . As initial state we should use the first data point from the validation set. Regarding this data point we approximately have q1=-0.0157 and q2= ...

Hi,
I just recognized that a template was uploaded (at least I think it is a template )

Update: However, matlab produces the following error if I call: mex DoublePendulum_C_....
Error using mex (line 206)
Unable to complete successfully.

I see a diference, too. However, my robot does not reach the target position with non of the approaches (the first link is alway pointing updards). I am also not sure how to compute xd. Its a litte strage if I use the Jacobian and qd to obtain xd, since I used this xd afterwards to obtain qd again :...

Kann sein, dass die Antwort nein ist. Zumindestens steht folgendes in dem Einfürhungsfoliensatz vom Computer Vision 1 (SoSe 2013):
Maschinelles Lernen - Statistische Verfahren II (Roth, WS, not in 2013)
(http://www.gris.tu-darmstadt.de/teachin ... -intro.pdf, Slide 11)

Wahrscheinlich nicht. In den Einführungsfoliensatz zu Computer Vision 1 (SoSe 2013) steht folgendes:
"Computer Vision II (Roth, WS, not in 2013)"

Ich glaube der Turnus soll gewechselt werden, so dass CV1 im WS liegt und CV2 im SoSe.

Hmm... I think the problem occurs because of my transformed image. It is probably wrong (see here: viewtopic.php?f=290&t=28497). Don't know if its just the size or the whole procedure. I think I have to spend sime time on it again ^^

Any other ideas of putting those images together? For me it's the same. I have no clue :D And I do not find any ressources that provide my some kind of clue. I tried template matching (but the images are just too large) and I used a variant were I calculated some kind of bounding box (using the NaN...

Hi, the task says "The rest part of the panorama image should be reconstructed from transformed a4p3b.png.". I wanted to ask what we are aspected to do (to get all points ^^) ? It is sufficient to implement a simple method which copies the transformend image into the panorama (so that it looks nice/...

Hi, my homography estimation is not very stable. Sometimes the resuling image looks very good, sometimes its really messed up. I follow the tipps from here https://www.fachschaft.informatik.tu-darmstadt.de/forum/viewtopic.php?f=290&t=28472 (threshold for chi square + use only the remaining pairs for...

lustiz hat geschrieben:

hymGo hat geschrieben:I hope that did not confuse you

It was a bit confusing, but I think it helped
Consequently, my result looks a bit more appealing (and is hopefully right) . Now, I need to reconstruct the whole panorama ^^

Yes, I thought that too. Trying it the other way around reluts in a bad results, as well
Maybee I have todo some re-thinking (even so I really thought my soultion should be right xD)

Hi, I just can't find my error, so I hope someone else can give me a hint :) I try to transform the image a4p3b into the plane of a4p3a. Therefore, I proceed as stated in the task. First, I use the homography to get the coordinates of a4p3b in the plane of a4p3a. Afterwards, I use interp2 to interpo...

C determines the degree of punishment for the slack variables. To achieve results as in a) (or rather no slack) C must be choosen ...
Sorry can't give the full answer. If you don't get it, just try it out. There are only two possibilities

As far as times are concerned, extracting the features takes most of the time (ca 1 minute for all images). KMeans takes about 4-6 seconds. All other computations are neglectable. Btw: My processor is a core 2 duo 2.5Ghz. I have nearly the same processor. Loading the images takes ca. 80 secs. and t...

I get a warning, too. Anyway, the displayed results seems to be correct. Warning: Trust-region-reflective algorithm does not solve this type of problem, using active-set algorithm. For more help, see Choosing the Algorithm in the documentation. > In quadprog at 371 In train_svm_s at 37 In problem1a ...