Diabetic foot ulcer (DFU) is a severe condition that requires close monitoring and management. Experienced clinical staff are best-positioned to provide ground truth annotations to train machine learning methods to auto-delineate the ulcer. In this paper, we revisit the largest DFU dataset for segmentation that was used in the Diabetic Foot Ulcer Grand Challenge 2022 (DFUC 2022) and provide further analysis. The DFUC 2022 segmentation dataset provides clinical expert delineation of ulcer regions. We assess whether the clinical delineations are easily machine interpretable by deep learning networks or if image-processing refined contours should be used. This paper demonstrates that image processing using refined delineations as ground truth can provide better agreement with machine-predicted results. With the in-depth understanding and observation of baseline models, we provide insights to guide future development in DFU segmentation. This dataset continues to serve as the challenge dataset for DFUC 2024, and it is available at: https://dfu-challenge.github.io/ .