Abstract : Face images captured by real-world video surveillance applications usually have low resolution. This leads to poor performance or even failure of most face recognition algorithms. As a consequence, identifying the face of the query in low resolution, based on the high-resolution image gallery, proves to be a huge challenge. To address this problem, a novel multi-resolution convolutional neural network (MRCNN) model is proposed in order to study the consistent feature representation from high-resolution and low-resolution face images. First, the corresponding labeled multi-resolution face images are utilized to train the MRCNN model. After this process, the trained model is used as the feature extractor in order to obtain features for the targets in the gallery and query images, respectively. Finally, the nearest neighbor method is applied as the classifier for the purpose of final identification. The experimental results from the two publicly available databases demonstrate the superiority of the proposed MRCNN.