ImageJ is good. However, far more important is that you load the samples at least 3 times so you get an idea about loading and blotting variation (technical replicas). It is also very important to do different exposures of your western blots and compare them all to detect non-linear responses. Ideally, if you want to compare the intensity of two samples and be really accurate, you do the western blot again with dilution series of the two samples you want to compare (loading 5, 10, 20, 35, and 50 micrograms total protein). If you get your result, you can scan all samples and plot a dose-response curve for the two smaples. Even then I would do 2-3 exposures and compare everything.

Note: Quantification of western blot signals is difficult when you are self-critical. It all depends on how accurate you want the method to be, and how different the signals are. If the differences are marginal, you need biological replicas as well, not just technical replicas.

there is a software called ImageJ. You acquire jpeg or tiff files into imageJ and then convert it to 8bit type inverted image (band becomes white and background black) and you can quantitate band comparing with a negative control lane.

Yes as Bahrat stated.IMAGEJ ProGRAM YOU CAN DOWNLOAD A FREE TRIAL FROM GOOGLE

As aforementioned, ImageJ is ideal for analyzing band intensity of western blot or agarose gel and ImageJ is an open source (free) software. Herewith, handout has been attached, u can read it and do yourself.

Thank you all, very helpful info. I will download it now

ImageJ is good. However, far more important is that you load the samples at least 3 times so you get an idea about loading and blotting variation (technical replicas). It is also very important to do different exposures of your western blots and compare them all to detect non-linear responses. Ideally, if you want to compare the intensity of two samples and be really accurate, you do the western blot again with dilution series of the two samples you want to compare (loading 5, 10, 20, 35, and 50 micrograms total protein). If you get your result, you can scan all samples and plot a dose-response curve for the two smaples. Even then I would do 2-3 exposures and compare everything.

Note: Quantification of western blot signals is difficult when you are self-critical. It all depends on how accurate you want the method to be, and how different the signals are. If the differences are marginal, you need biological replicas as well, not just technical replicas.

Two very simple and clear guides:

http://lukemiller.org/index.php/2010/11/analyzing-gels-and-western-blots-with-image-j/

http://www.di.uq.edu.au/sparqimagejblots

ImageJ is for free.

Very important considerations before starting with the experiment: load the same amount of total protein in each well of the gel. Control that the transfer efficiency is equal allover the gel (stain the gel with Comassie after transfer). When acquiring the digital image be sure that the signal is not saturated and save the pictures in TIFF uncompressed format. NEVER use jpeg format for quantification, it is a lossy compression algorithm and therefore you lose information every time you save the file.

Technical and especially biological replicas (at least 3) are important for statistical significance. 

I Agree with previous comments. ImageJ is free and excellent tutorial about how to use it for western blot quantification is available-http://lukemiller.org/index.php/2010/11/analyzing-gels-and-western-blots-with-image-j/

However, there's  one drawback with ImageJ. ImageJ cannot quantify unequal/irregular bands (with Analyze-Gel- menu).  Your bands should be of approximately equal size and they should be at the same level (no zig zag pattern).

I tried it after I read the instruction. The tutorial is excellent, and its easy to quantify by imageJ

Thanks to all

Unless you have a Licor Odyssey Scanner and Image Studio software!! It truly takes Western Blotting to another level! 

Yes, ImageJ is an excellent tool for quantification of band intensity in immunoblotting and it is a free software

Hello, I was wondering if ImageJ or any other software can quantify densities of bands that aren't on the same horizontal or vertical line ? i need to semi-quantify bands of my western blot that are distributed on different levels on the blot,, is there a way to do this ? Any thoughts will be appreciated