Identifying elephant photos by multi-curve matching

摘要: We present in this paper an elephant photo identification system based on the shape comparison of nicks characterizing elephants' ears. The method we propose can deal with very cluttered and noisy images as ones commonly used by zoologists for wild identification. Difficult segmentation problems are solved using rough position information input user. Such is a basis set normalization hypotheses aiming at comparing query Q different photos system's database possibly representing same individual Q. proposed method, matching multiple, non-connected curves, be applied to retrieval problems. Examples real shown.