MATLAB for image segmentation

[XTRNTR]

Sorry, i don't know if this is the right place to post this thread?

I will be using MATLAB for my FYP on using techniques such as neural network (kohonen self organizing map) or discrete wavelet transforms for the application of image segmentation. Does any of these terms ring a bell?

The project is basically for me to experiment with these techniques, compare their efficiency, shortcomings and maybe even combine both approaches. I took this FYP despite not having any background knowledge in the field of image processing because i wanted to learn something new, anyway i would appreciate it if anyone is up for discussing these with me. I don't expect to be spoon fed answers here.

Thanks for reading!

 

[davidktw]

Sorry, i don't know if this is the right place to post this thread?

I will be using MATLAB for my FYP on using techniques such as neural network (kohonen self organizing map) or discrete wavelet transforms for the application of image segmentation. Does any of these terms ring a bell?

The project is basically for me to experiment with these techniques, compare their efficiency, shortcomings and maybe even combine both approaches. I took this FYP despite not having any background knowledge in the field of image processing because i wanted to learn something new, anyway i would appreciate it if anyone is up for discussing these with me. I don't expect to be spoon fed answers here.

Thanks for reading!

I highly recommend that you change your FYP to something you at least have some basic ideas about. FYP while it is something to encourage research and development, I personally don't feel you should embark on subjects you don't even have the slightest idea of what it's about. It's not about the grade that you will get at the end of the day, it's whether you will delivery timely.

If you can't deliver, it's going to really affect your grades badly.

Neural network is a weighage and feedback mechanism that forms the idea of a brain like analytical engine. I'm not expert in AI, but have learnt about such mechanism during my university. It's suppose to operate as a self tuning mechanism with the capacity to adapt. You will need to train the neural network to get the values you wanted. Eventually hopefully you will get a tuned network that can operate in scenarios outside of the training material.

Discrete wavelet transform is the algorithm used in JPEG2000. That's as far as I know. I don't study this particular algorithm in depth except DCT used in JPEG. It is suppose to be able to store the same image with similar quality with much lower bitrate versus one that will be stored using DCT such as JPEG without those high frequency artifacts made. Refer to Why Do Math? for more information