We propose and demonstrate a technique based on digital holography for three-dimensional (3D) object recognition. In our method, the complex wavefront of the 3D object to be recognized is recorded as a Fresnel hologram using a CCD camera. A two step numerical reconstruction process, retrieval of complex wavefront recorded on the CCD plane and propagation of wavefront to the object plane, gives a complex valued two-dimensional (2D) image, which contains information regarding the amplitude and phase of the object wavefront in a 2D plane. The encoding of 3D object information as 2D digital image enables the use of various image-processing techniques for the post processing of the data. We use a Mexican hat wavelet matched filter (WMF) to discriminate between two different objects. WMF performs wavelet transform (WT) to enhance significant features of the images and the correlation of the so obtained WT coefficients in a single step. Digital wavelet matched filtering improves the discrimination capability and results in sharper correlation peaks as compared to classical matched filtering. As compared to many of the 3D recognition techniques based on the processing of 2D perspectives, our system has lesser complexity at the implementation and operational levels.