实际上，域内相机样式差异是人员重新ID的关键影响因素，因为在测试过程中

In fact, the differences within the camera style is the key factor affecting the reintegration of ID, because during the test, matching query and its basic truth is captured by a different camera. <br><br><br>If you do not consider the image changes within fine-grained, then set on the source of the transfer of training learning model will probably only capture the overall data deviation between the two domains, and have greater focus image changes within the target domain will test problem appear

In fact, the difference in camera style within the domain is a key factor in the re-ID of a person, because during testing, the query and its basic true match are captured by different cameras.<br><br>If you do not consider fine-grained intra-domain image changes, the transfer learning model trained on the source set will likely capture only the overall data deviation between the two domains and will experience problems when encountering large rand in the target domain test set

In fact, camera style differences within the domain are the key influencing factors of personnel re ID, because during the test, the query and its basic truth matching are captured by different cameras.<br>If the fine-grained intra domain image changes are not considered, the transfer learning model trained on the source set may only capture the overall data deviation between the two domains, and there will be problems when large intra domain image changes are encountered in the target domain test set<br>