Specifically, the proposed fusion method mainly comprises three crucial parts: aspect extraction, aspect fusion and graphic reconstruction. We For starters benefit from seven convolutional blocks to extract the impression attributes from source photos. Then, the extracted convolutional attributes are fused via the proposed fusion method inside the element fusion layer. https://erniep604rye5.bmswiki.com/user