Gaussian low-pass filtering is a common post-process operation which is exploited to blur and conceal these discontinuities at the border of tampered objects introduced by copy & paste operation, making the tampered image more realistic. In this paper, a novel approach for detecting Gaussian low-pass filtering in digital images is proposed based on the frequency residual function which presents different attribute in median-high frequency domain between the original image and the filtered image. A Gaussian low-pass filter bank is utilized to obtain a series of frequency residual functions and a bandwidth feature vector is proposed to identify the filtered image with the Support Vector Machine (SVM). Excellent experimental results verify the effectiveness of our proposed approach.Gaussian low-pass filtering is a common post-process operation which is exploited to blur and conceal these discontinuities at the border of tampered objects introduced by copy & paste operation, making the tampered image more realistic. In this paper, a novel approach for detecting Gaussian low-pass filtering in digital images is proposed based on the frequency residual function which presents different attribute in median-high frequency domain between the original image and the filtered image. A Gaussian low-pass filter bank is utilized to obtain a series of frequency residual functions and a bandwidth feature vector is proposed to identify the filtered image with the Support Vector Machine (SVM). Excellent experimental results verify the effectiveness of our proposed approach.