What are the fields in Bioinformatics related to Image processing or machine vision except Microarrays?

Question

: I want to work on such fields in bioinformatics that use Image processing or Machine Vision directly (except Micrroarrays) as my MSc thesis.

Answer 1

In almost every field of biology and chemistry there are huge numbers of machines that collect large amounts of 2D, 3D, and even 4D+ data. Things that collect images include microarrays, crystallography, NMR (after processing), electron microscopy, and many other techniques. In electron microscopy, specifically, image processing is the "rate-limiting step." Basically, the way electron microscopy works is that you take many (thousands, millions, billions) of images of complex macromolecular assemblies on a surface. They are oriented randomly (e.g. rotation in x, y, and z are random), and so automated programs need to start collecting the images and classify which ones look similar, assume they have similar x/y/z rotation values, generate a 3D structure based on that, and then narrow down images to more specific x/y/z rotation values. The more images one collects the better an overall structure of the macromolecular process is obtained. However, the image processing is one of the biggest problems in EM, dealing with symmetry, classifying, checking for errors and biases in the bootstrapping process, etc.

If you want to talk about an even bigger problem, consider the processing and automated assignment of NMR (nuclear magnetic resonance) data of proteins. There you have a sequence, and therefore an exact known chemical structure, and each peak in the spectra corresponds either to an atom or a conection between the atoms. In order to generate a 3D structure of the protein you need to assign essentially every single peak in the spectra (which are 2D, 3D, and 4D). Although there are automated programs out there, their reliability is low and so people use manual or semi-automated processes. Yet again, another big field where image processing is really important. You can check out the wikipedia articles on EM http://en.wikipedia.org/wiki/Cryo-electron_microscopy (not such a good article) and protein NMR http://en.wikipedia.org/wiki/Protein_nmr (good article) to get a slightly better idea of some of these techniques.

MRI, CT, PET and all the other medical imaging techniques also generate huge amounts of date, and are more complicated because of movement (heart beating, blood flow, patient movement, etc.). As soon as you put the word medical in anything, of course, more people care about it and more money's available. The NIH has many may many inter-disciplinary imaging components in multiple institutes in case you're wondering about possibilities after you graduate.

Good luck.

Answer 2

i like how no person has any actual answer -- they only seek for to alter the question to greater healthful them. some Creationist Advances: imaginative Writing -- it must be VERY imaginative Tortured good judgment Flood Geology (see 2 above entries) Selective historic previous -- Noah's large Flood occurred on the same time through fact the 1st large Egyptian Dynasties, yet controlled to spare the Egyptians. (to not point out the chinese language) better recommendations Washing of the Gullible

Answer 3

CAT scanners and similar medical equipment.

Modern sequencers rely on image procesing.

2-D gels.

Probably also some applications for asisting patologists and radiologist and acessing microspope images and X-ray images, respectively, but I don't know about those fields.

Answer 4

There is an example for dna modeling and image processing:
Please Google: Morbid Motion Monitor

or

http://webcache.googleusercontent.com/search?q=cache:5TmYI-Igz2YJ:https://poseidon.cs.abo.fi/trac/project_course2012/raw-attachment/wiki/WikiStart/MorbidMotionMonitor.pdf+&cd=1&hl=en&ct=clnk&gl=fi&client=firefox-a