In this paper, we present a method of camera-based document image retrieval for heterogeneous-content document using a multi-layer separating approach. We use Locally Likely Arrangement Hashing (LLAH) extracting text features on the layer which contains text. In addition, we employ a technique of reducing the memory required for storing the hash table. Experiment result show that our approach is efficient in term of accuracy result and real-time retrieval for heterogeneous-content document camera-based retrieval.