Alignment of core point for shoeprint analysis and retrieval

Abstract

The purpose of this study is to propose a shoeprint retrieval method based on core point alignment for pattern analysis. The proposed method firstly detects contour points in a black-and-white shoeprint image. Those reliable contour points are selected to simulate the left and right sidelines of the shoeprint by curve fitting method. Subsequently, the most concave points along the left and right sidelines can determine the core point of the shoeprint, thereby partitioning the shoeprint into circular regions. Next, the Zernike moments of the circular regions are calculated for pattern description of each region. Finally, the Euclidean distance is measured to match the shoeprints with the same pattern. The highest = 0.726 is obtained from the first four Zernike moments with the radius = 90 pixels and three baselines. The experimental results also show that Zernike method in any orders always outperforms the compared moment invariants and GLCM method. This study has verified that the proposed method can effectively align the shoeprints for pattern comparison.

Quantifying the similarity of 2D images using edge pixels: an application to the forensic comparison of footwear impressions

“…Bouridane et al [4] proposed a fully automatic system for matching and retrieval of 2D images that uses a fractal decomposition based on pattern comparison. AlGarni and Hamiane [1] extract features of the outsole using Hu moments [15] in their algorithm and Wei and Gwo [34] and Gwo and Wei [13] summarize outsole patterns using Zernike moments. Patil and Kulkarni [24] propose a matching method that relies on a Gabor transform for feature extraction and a Radon transform for estimating the rotation angle between the images.…”

Crime Scene Shoeprint Retrieval Using Hybrid Features and Neighboring Images

“…They can handle high quality shoeprint images; however, this method is sensitive to variations in rotations and translations. The moment invariant features were used for shoeprint retrieval [2,3], and they can work well for complete shoeprints; however, partial shoeprints are not considered. Chazal et al [4] and Gueham et al [5,6] used the Fourier transform to analyze the frequency spectra of shoeprint images, but the methods are sensitive to partial shoeprints.…”

“…Most of the existing approaches focus on designing low-level features to highlight the most similar aspects of shoeprints. From the perspective of methodology, most of these shoeprint retrieval methods fall into two categories: (i) retrieval of shoeprints using holistic features [1][2][3][4][5][6][7][8][9][10][11][12][13]; and (ii) retrieval of shoeprints using local features [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29]. Although both kinds of methods are able to search for shoeprint images with visually similar shoe patterns, their retrieval precision may vary dramatically among low quality crime scene shoeprint images.…”

LOSGSR: Learned Opinion Score Guided Shoeprint Retrieval

“…However, this method is sensitive to variations in rotations and translations. The moments invariant features are used for automatic shoeprint retrieval, such as Hu's moments [3] and Zernike moments [4], [5], and they report good results of their methods; however, the shoeprints that they used are clear and complete; partial data is not considered. Patil and Kulkarni [6] provide a retrieval method that uses the Radon transform and the Gabor feature to align and represent a shoeprint, and the method shows good performance for partial shoeprints generated from full shoeprints in a rigorous manner.…”