Linkoping University, Department of Biomedical Engineering, 58283, Linkoping, Sweden.
Linkoping University Hospital, Linkoping Burn Centre, 58185, Linkoping, Sweden.
Sci Rep. 2017 Dec 1;7(1):16744. doi: 10.1038/s41598-017-16914-0.
Assessment of burn scars is an important study in both medical research and clinical settings because it can help determine response to burn treatment and plan optimal surgical procedures. Scar rating has been performed using both subjective observations and objective measuring devices. However, there is still a lack of consensus with respect to the accuracy, reproducibility, and feasibility of the current methods. Computerized scar assessment appears to have potential for meeting such requirements but has been rarely found in literature. In this paper an image analysis and pattern classification approach for automating burn scar rating based on the Vancouver Scar Scale (VSS) was developed. Using the image data of pediatric patients, a rating accuracy of 85% was obtained, while 92% and 98% were achieved for the tolerances of one VSS score and two VSS scores, respectively. The experimental results suggest that the proposed approach is very promising as a tool for clinical burn scar assessment that is reproducible and cost-effective.
烧伤瘢痕评估在医学研究和临床环境中都是一项重要的研究，因为它可以帮助确定烧伤治疗的反应并规划最佳的手术程序。瘢痕评分既可以通过主观观察，也可以通过客观测量设备进行。然而，目前的方法在准确性、可重复性和可行性方面仍然缺乏共识。计算机化的瘢痕评估似乎有可能满足这些要求，但在文献中很少见。本文提出了一种基于温哥华瘢痕量表（VSS）的自动烧伤瘢痕评分的图像分析和模式分类方法。使用儿科患者的图像数据，获得了 85%的评分准确率，而对于 VSS 评分的容差为 1 分和 2 分的情况，分别达到了 92%和 98%的准确率。实验结果表明，该方法作为一种可重复、经济有效的临床烧伤瘢痕评估工具具有很大的应用前景。
