An approach to the wavelet multiresolution analysis is realized using Coiflet wavelets and a two-wave mixing arrangement via transmission gratings in a bismuth silicon oxide (BSO) crystal. A Coiflet filter acts not only as a band-pass filter but as a high-pass or low-pass filter. 2-D wavelet bases are generated by a tensorial product of the 1-D Coiflet bases and are separable in x and y. Therefore, the 1-D band-pass property produces direction ability in the 2-D tensorial products. In this way, unlike the Mexican hat-based filters, the Coiflet-based filters allow selecting horizontal, vertical or diagonals details of an original image. The decomposition of a signal in two levels of multiresolution is experimentally demonstrated in this scheme. The original image is discomposed in an average image and several detail images depending on the level of multiresolution. Coif12 wavelets are used to achieve the multiresolution analysis. The photorefractive implementation of the wavelet analysis is numerically simulated and experimentally performed. An agreement between numerical and experimental results is obtained.