Assuming you have good quality images, with enough resolution, then counting nuclei is the most straightforward way to count cells. The steps are:
1. Duplicate the image.
2. Optional: Apply a 2 pixel median filter to the duplicate to smooth it ("Process", "Filters")
3. Apply a threshold to the filtered duplicate ("Image", "Adjust", "Threshold"). Make sure "Background" is set to "black." You want automatic thresholding. Test several algorithms. Default or Huang often work well, but try them all until you find one that best selects your label. Apply the threshold to create a binary image.
4. Use "Analyze", "Analyze Particles" to create ROIs of the structures in the binary image. You can impose size, circularity restrictions here and also exclude things on the edge which I usually do. Once the objects are identified, hit "t" to send them to the ROI manager.
5. Select your raw image and apply the ROIs from the manager onto that image.
6. Under "Analyze", "Set Measurements" you can select which parameters to measure. if you only want number, then this really doesn't matter. In the ROI manager click "Measure" and a window will open with a list of all structures.
This is a protocol explaining how to do manual or automated cell counting. You will need the plugin Cell counter, which you can downoad for free, or if you install Fiji, it's included in the software. Good luck!
There are many software such as Fiji , ImageJ and Image Pro Plus .
I used Image Pro Plus software with IHC , Its depends on Integrated Optical Density (IOD): which mean the reports the average intensity/density of each object. This value will be expressed in terms of the current intensity/density mode and calibration."
The best explanation I have found: IOD = area x average density