This paper proposes an image-based technique for shredded document reconstruction. The problem is different from solving jigsaw puzzles since curved boundaries and color information are not available. Currently most research on document recovery focuses on image feature exaction and analysis. In this work, we present a complete procedure which is capable of reconstructing a full page of shredded document. Similarity measure based on shred boundary correlation is defined for pattern matching. A weighted digraph is then used to derive the final shred sorting result. Experiments are presented for both the synthetic and real datasets.