The present invention discloses a finger vein feature comparison method and device, a storage medium and a processor. The method comprises the steps of pre-processing an inputted finger vein image; adopting a Gabor function to process the pre-processed image to obtain the Gabor vector features; carrying out the dimensionality reduction processing on the obtained Gabor vector features; carrying outthe binary Gabor gradient feature extraction on the Gabor vector features after the dimensionality reduction; calculating the gradients of the Gabor coefficients in the corresponding directions underdifferent frequency parameters, processing the gradient values into the values of 0 and 1 in a binary manner according to the sign symbols of the gradient values, and then placing the corresponding values in the brand new vectors corresponding to the component positions of the block coordinate, frequency and direction parameter indexes, and generating the binary expressed brand new vector features; comparing the binary expressed brand new vector features and outputting the comparison results. According to the present invention, the stable expression features are obtained, and the performanceof a finger vein identification algorithm is improved.