Evaluation of 3D Measuring Methods for Body Surface Damage and Scars

Objective To assess the accuracy and reliability of the two 3D measuring methods, structured-light scanning and single-camera photogrammetry in forensic body surface damage and scar measurement. Methods The measuring results of 86 injuries by the ruler method, structured-light scanning and single-camera photogrammetry were compared and evaluated. The area measuring results of 13 simulated scars by the structured-light scanning, single-camera photogrammetry, Photoshop lasso pixel method and PDF reading software were compared and evaluated. The time consumed was recorded. The known specification information of the stickers was used as the standard value to compare the measuring accuracy of the ruler method, structural-light scanning and single-camera photogrammetry, and to calculate the root mean square error （RMSE）. The consistency evaluation of intraclass correlation coefficient （ICC） for the repeatability of 3 measuring results obtained by different operators and the same operator. Results The differences in results of the two 3D measuring methods and the ruler method had no statistical significance; the differences between measuring results made by the structured-light scanning, single-camera photogrammetry and PDF reading software and that of the Photoshop lasso pixel method had no statistical significance. The post-processing of the single-camera photogrammetry consumed more time than that of other methods. When the long-distance group （10-40 cm） was measured, the results obtained by the ruler method were shorter than the standard value. The RMSE value results were structured-light scanning < single-camera photogrammetry < ruler method. The ICC value intragroup and intergroup were greater than 0.99. Conclusion Both structured-light scanning and single-camera photogrammetry can be applied in recording and measuring forensic body surface damage. The former has better performance in measurement accuracy and stability, while the latter has better color performance but longer post-processing time.