Deep Binary Representation for Efficient Image Retrieval

Figure 2

The CNN network of our proposal. The upper part is the DBR network, and the lower one is the DBR-v3 network based on inception-v3 net. First, target hash code set is generated based on hash length and image category number . Then deep network is trained with raw images and target hash code. Finally, image retrieval is processed with the hash function, which is the trained network concatenated with a function.