<p>Track stiffness is a main parameter influencing the behavior of railway tracks transition zones due to sudden stiffness change. There are several developed techniques for track stiffness measurement including stationary and moving methods. In recent studies, Digital Image Correlation (DIC)-based methods are investigated as a way of deflection measurement of railway track components to be substituted for traditional ones. Among these, the investigation of track vertical deflection is less paid attention for the purpose of stiffness determination. In the present paper, two different pixel detection algorithms of Pixel Following (PF) and Darkest/Lightest Pixel Detection (DLPD) were developed to assess the time-dependent pixel movement during train loading. For this purpose, an approach embankment of a concrete bridge in Tehran-Hamedan route of Iranian railway track has been selected for field measurements. The results of statistical survey show that DLPD data have about 42% overestimation comparing to Linear Variable Differential Transformer (LVDT) data, which results in rejection of mentioned method for further calculations. But, PF method is able to detect deflection time histories with more than 90% agreement in normal conditions, validated by LVDT, as well as determination of peaks in the case of existing hanging sleeper and windy condition. The frequency contents of PF data also match to the train wheel/bogie configuration, meaning that there is no missing dominant frequency in the method. The calculated track stiffness values has a close approximate to those determined in previous studies, which can be employed in further railway track investigations as a novel technique.</p>