ORIGINAL ARTICLE 3D-based reconstruction using growing neural gas landmark: application to rapid prototyping in shoe last manufacturing Antonio Jimeno-Morenilla & Jose García-Rodriguez & Sergio Orts-Escolano & Miguel Davia-Aracil Received: 31 January 2013 / Accepted: 7 May 2013 # Springer-Verlag London 2013 Abstract Customizing shoe manufacturing is one of the great challenges in the footwear industry. It is a production model change where design adopts not only the main role, but also the main bottleneck. It is therefore necessary to accelerate this process by improving the accuracy of current methods. Rapid prototyping techniques are based on the reuse of manufactured footwear lasts so that they can be modified with CAD systems leading rapidly to new shoe models. In this work, we present a shoe last fast reconstruc- tion method that fits current design and manufacturing pro- cesses. The method is based on the scanning of shoe last obtaining sections and establishing a fixed number of land- marks onto those sections to reconstruct the shoe last 3D surface. Automated landmark extraction is accomplished through the use of the self-organizing network, the growing neural gas (GNG), which is able to topographically map the low dimensionality of the network to the high dimensional- ity of the contour manifold without requiring a priori knowl- edge of the input space structure. Moreover, our GNG landmark method is tolerant to noise and eliminates outliers. Our method accelerates up to 12 times the surface recon- struction and filtering processes used by the current shoe last design software. The proposed method offers higher accu- racy compared with methods with similar efficiency as voxel grid. Keywords Shoe manufacturing . Shoe last rapid prototyping . 3D surface reconstruction . Landmarking . Growing neural gas . Voxel grid 1 Introduction The shoe last design is the basis to perform the design of a footwear. This is an industrial object that must meet de- manding criteria in order to achieve a successful final prod- uct. The shoe last can be divided into two parts: the body and the toe. The body (see Fig. 1) is the part that extends from the base of the metatarsal to the ankle, and it is fundamental for a correct adjustment of the foot in the final product. The body should be fitted in the average population foot size, and therefore, the accuracy in the design and manufacture of this part must be very high (in the sector, it is typified in ±0.5 mm). The other part, the toe, defined from the metatarsals, has more freedom in terms of design, and most of the changes imposed by the current fashion trends lie on it. Figure 2 shows various types of shoe last tip. In [1], a method for analyzing the shoe last fitting to a particular foot, by checking the number of sections obtained from both the foot and the shoe last, is proposed. Factories do not dispose the old shoe lasts. This is be- cause a shoe last with a “good body” can lead to new designs that continue meeting the demanding ergonomic criteria. Therefore, the shoe lasts are reused, physically separating the toe from the body and molding a new toe on the previous body. This process is slow and complex. With the advent of computer-aided design (CAD)/computer- aided manufacturing (CAM) and rapid acquisition devices, it is possible to digitize old raised shoe lasts for reusing them in the shoe last design software. The first scanners were mechanical copiers (see Fig. 3), which are slow and inflexible, but very accurate and robust, and also of- fered a spatial ordered 3D information as shaped sec- tions that were easily treated by the CAD software. The main problem was the slowness that makes them not suitable for rapid prototyping environments, but yet still used in the industry [2]. A. Jimeno-Morenilla (*) : J. García-Rodriguez : S. Orts-Escolano : M. Davia-Aracil Department of Computer Technology, University of Alicante, San Vicente del Raspeig s/n, 03690 Alicante, Spain e-mail: jimeno@dtic.ua.es Int J Adv Manuf Technol DOI 10.1007/s00170-013-5061-3