I have the same doubt as he did, but in my scenario, I Created four different classes and only 2 had trained. one class has the same structure but another one is different in the structure to each item, so after the training it was working well in trained images but when I tried to do object detection in new images there was a huge of confusion in labeling a correct class object. Is object detection only apply on defined structured items? or is there any other object detection model that use full for this scenerio. currently im using yolov11