The invention provides an in-vivo detection method in a finger vein recognition process. The method is characterized by comprising the following steps of: acquiring certain frames of finger vein video images in the finger vein recognition process; selecting partial areas with finger veins as in-vivo vein detection areas; and finally carrying out vein line slight inflation and contraction movement detection by using an Euler video movement amplification algorithm, and determining whether the finger veins are in-vivo finger veins or not according to the detection result. The method has the effects of no needing to increase hardware of an original finger vein recognition platform, keeping the advantages of non-contact acquisition and being applied to in-vivo detection of hand back and palm veins.