E-Cast: Fast end-to-end 3D scanning and printing system of orthopedic casts
The main challenge of the E-Cast project is to automate and renew the medical treatment of fractures or tendinopathies. The traditional treatment is based on the immobilization of the limb through a plaster cast. Such a cast, based on a 1851 technology, is heavy, not breathable, doesn’t allow patients to wash, gets dirty easily, causes swelling and numbness, doesn’t allow to play sports, it’s aesthetically unpleasant and can cause allergic reactions. Furthermore, the production of plaster cast requires highly qualified personnel and, being a handmade product, patients often have to return to a healthcare facility to reposition, modify and repair the plaster.
The E-Cast system exploits cutting-edge computer vision technologies and patented 3D printing technologies to produce new generation plastic casts in a fully automatic and accurate way. The 3D printed casts built with the E-Cast system offer several advantages over traditional plaster casts, both for doctors and patients. They are washable, suitable for sports, breathable, light (5 times lighter than a traditional plaster cast) and, thanks to the accuracy of 3D scanning and automated production, they offer a perfect fit. The E-Cast system includes a state-of-the-art, customized 3D scanner to recover the structure of the body part of interest. The scanning process is instantaneous and fully automated, while the obtained 3D reconstruction is very accurate and in high resolution. Thanks to an easy to use application, the orthopaedic cast can be printed in 3D in minutes, making the system quick and end-to-end, from click to 3D printing.
The E-Cast project has reached substantial achievements during the DIGI-B-CUBE financing period. In particular, the E-Cast front-end, the 3D scanner and mesh reconstruction part, has been improved and the system is now capable of reproducing high-fidelity (sub millimeter accuracy) 3D reconstruction of human limbs in less than 2 minutes by exploiting state-of-the-art multi-view stereo matching techniques. An innovative approach that involves Deep Learning for the segmentation of the 16 images that are retrieved by the 3D scanner has been studied and is under development in the last part of the project, likewise a novel 3D reconstruction based on Deep Convolutional Networks is going to be introduced in the main pipeline as initial guess for the stereo-matching pipeline. The cast synthesis has been further developed, this speeds up the production of the 3D printed cast and make the whole system easily usable also by non-expert users.
The adoption of 3D printed casts in the orthopaedic threatments is very important and has a huge impact first of all on the patients, for example, this novel cast is radiotransparent and allows healthcare professionals to monitor progresses through X-rays or immediately initiate magnetic therapies. Moreover, It allows to reduce the overload of first aid facilities, it does not require qualified personnel, is cheaper than a traditional system and is made of 100% recyclable plastic.