Wavelet based Contour Extraction for Person Identification in Dental Images

Forensic Odontology is a specialised field of victim identification that includes the identification of salvaged humans as well as the proof of identity including whole or scrappy bodies. Contour information of dental x ray images provides better matching for person identification. In this paper, we propose an automatic Wavelet Sliding Neighborhood (WSN) method for human forensic identification which comprises of preprocessing, integral intensity projection, contour extraction and shape matching. The first step preprocessing is used to remove the noise present in the dental x ray or radiographic images. The second step entangles integral intensity projection for untying the lower jaw, upper jaw and individual teeth. Third stage uses wavelet based sliding neighborhood to extract the shape of the dental image without any discontinuities in the edges of an image. The final stage, shape matching is used for estimating the distance measure for matching ante-mortem (AM) and post-mortem (PM) dental records. The experimental results demonstrate that our proposed WSN method outperforms and achieves a highest similarity than the existing approaches with a computation time of 1.25s.