Mohamed Alzaki's profile - activity

@StevenPuttemans Any thoughts?

I am using haar cascade to detect my target objects and it does the job very well in my case 70% overall detection rate.

Here is an example of the detected objects:

Now i'd like to improve the detection such that it doesn't only detect the target objects region as a rectangle, i'd like to detect it accurately by its shape, is SURF is what i am looking for? "Given that the objects' shapes scope is wide ".

Please take a look of my positive samples album to get an idea of what i'm dealing with.

Here is an album of the positive samples.

Target objects in the images are the kidneys

Positive Samples Album

And if so can i use SURF in conjunction with Haar cascade, is that a good practice "Would improve the accuracy?" or it would only slow down the process?

Thanks for your answer that was straightforward and helpful answer.

Another thing does increasing the dimensions of samples improve the accuracy ? detection rate is like 70% now what do you suggest to improve it.

I'd like to increase the amount of memory used in the training using opencv_traincascade, i tried to add precalcValBufSize and precalcIdxBufSize parameters but no thing seems to be changed my machine is 64GB ram and running ubuntu OS when i see the consumption of training its always around 7.5 GB no changed occurred after adding the parameters although i see PrecalcValBufSize and precalcIdxBufSize in the terminal has been set to a value before it enters the stage. Any thoughts?

I guess its hanged there because HitRate doesn't increase at all for more than a day now but FalseAlarm does change, i think its maybe because i didn't set sufficient parameters? or i am wrong, in some cases it takes too long with the same numbers.

24 hours passed in the same stage with exactly same SMP and same HR but the FA decreased a little bit I've updated the answer with the command i use for haar training for the second matter.

@StevenPuttemans i am very interested to hear your opinion in this matter.

@berak updated please check the images thanks for your note.

I am working on a project that should take a ct scan image as an input and it should detect the kidney and do some processing on it, my question is due to the low features in that type of images and variant types of kidneys, is haar classifier qualified to detect such object?

The objects in the red rectangles (kidneys) are the target objects which i want to detect.

Another question i am training with the samples at the time i am writing this post, with 579 positive images and around 1k negative images i am using haartraining and it is running for 14 hours now and is in stage 4 but i've noticed in that stage that the SMP is always around 0% - 2% is that normal?

Here is the parameters i use for haartraining

haartraining.exe -data cascades -vec vector/kidney.vec -bg negative/bg.txt -npos 579 -nneg 973 -nstages 15 -mem 1024 -mode ALL -w 24 -h 24

rem -nonsym