In your case it would seem that you did not get any amplification in those wells. If you look at the Y-axis, the scale is far smaller than would be expected. Compare them to the values in your housekeeping gene, which should have a much larger peak.
The melt curve after qPCR is sort of comparable to running out samples on a gel. As John Hardy Lockhart pointed out, it looks like you have no amplification of any product in 3 of 4 samples and only a tiny bit of amplification in 1 sample.
But the melt-curve is the LEAST informative part of a qPCR experiment. What do your Ct values & standard curve look like?