A method for finding high accuracy surface zones on 3D printed bone models.

The use of three-dimensional (3D) printing for surgical applications is steadily increasing. Errors in the printed models can lead to complications, especially when the model is used for surgery planning or diagnostics. In patient care, the validation of printed models should therefore be performed routinely. However, there currently is no standard method to determine whether the printed model meets the necessary quality requirements. In this work, we present a method that not only finds surface deviations of a printed model, but also shows high accuracy zones of a potentially corrupted model, that are safe to be used for surgery planning. Our method was tested on printed patient bone models with acetabular fractures and was compared to two common methods in orthopedics, simple landmark registration as well as landmark plus subsequent iterative closest point registration. In order to find suitable parameters and to evaluate the performance of our method, 15 digital acetabular bone models were artificially deformed, imitating four typical 3D printing errors. A sensitivity of over 95% and a specificity of over 99% was observed in finding these surface deformations. Then, the method was applied to 32 printed models that had been re-digitized using a computed tomography scanner. It was found that only 25% of these printed models were free of significant deformations. However, focussing on two common implant locations, our method revealed that 72% of the models were within the acceptable error tolerance. In comparison, simple landmark registration resulted in a 9% acceptance rate and landmark registration followed by iterative closest point registration resulted in a 41% acceptance rate. This outcome shows that our method, named Similarity Subgroups Registration, allows clinicians to safely use partially corrupted 3D printed models for surgery planning. This improves efficiency and reduces time to treatment by avoiding reprints. The similarity subgroups registration is applicable in further clinical domains as well as non-medical applications that share the requirement of local high accuracy zones on the surface of a 3D model.

AggNet: Deep Learning From Crowds for Mitosis Detection in Breast Cancer Histology Images.

The lack of publicly available ground-truth data has been identified as the major challenge for transferring recent developments in deep learning to the biomedical imaging domain. Though crowdsourcing has enabled annotation of large scale databases for real world images, its application for biomedical purposes requires a deeper understanding and hence, more precise definition of the actual annotation task. The fact that expert tasks are being outsourced to non-expert users may lead to noisy annotations introducing disagreement between users. Despite being a valuable resource for learning annotation models from crowdsourcing, conventional machine-learning methods may have difficulties dealing with noisy annotations during training. In this manuscript, we present a new concept for learning from crowds that handle data aggregation directly as part of the learning process of the convolutional neural network (CNN) via additional crowdsourcing layer (AggNet). Besides, we present an experimental study on learning from crowds designed to answer the following questions. 1) Can deep CNN be trained with data collected from crowdsourcing? 2) How to adapt the CNN to train on multiple types of annotation datasets (ground truth and crowd-based)? 3) How does the choice of annotation and aggregation affect the accuracy? Our experimental setup involved Annot8, a self-implemented web-platform based on Crowdflower API realizing image annotation tasks for a publicly available biomedical image database. Our results give valuable insights into the functionality of deep CNN learning from crowd annotations and prove the necessity of data aggregation integration.

NeuroKinect: A Novel Low-Cost 3Dvideo-EEG System for Epileptic Seizure Motion Quantification.

Epilepsy is a common neurological disorder which affects 0.5-1% of the world population. Its diagnosis relies both on Electroencephalogram (EEG) findings and characteristic seizure-induced body movements--called seizure semiology. Thus, synchronous EEG and (2D)video recording systems (known as Video-EEG) are the most accurate tools for epilepsy diagnosis. Despite the establishment of several quantitative methods for EEG analysis, seizure semiology is still analyzed by visual inspection, based on epileptologists' subjective interpretation of the movements of interest (MOIs) that occur during recorded seizures. In this contribution, we present NeuroKinect, a low-cost, easy to setup and operate solution for a novel 3Dvideo-EEG system. It is based on a RGB-D sensor (Microsoft Kinect camera) and performs 24/7 monitoring of an Epilepsy Monitoring Unit (EMU) bed. It does not require the attachment of any reflectors or sensors to the patient's body and has a very low maintenance load. To evaluate its performance and usability, we mounted a state-of-the-art 6-camera motion-capture system and our low-cost solution over the same EMU bed. A comparative study of seizure-simulated MOIs showed an average correlation of the resulting 3D motion trajectories of 84.2%. Then, we used our system on the routine of an EMU and collected 9 different seizures where we could perform 3D kinematic analysis of 42 MOIs arising from the temporal (TLE) (n = 19) and extratemporal (ETE) brain regions (n = 23). The obtained results showed that movement displacement and movement extent discriminated both seizure MOI groups with statistically significant levels (mean = 0.15 m vs. 0.44 m, p<0.001; mean = 0.068 m(3) vs. 0.14 m(3), p<0.05, respectively). Furthermore, TLE MOIs were significantly shorter than ETE (mean = 23 seconds vs 35 seconds, p<0.01) and presented higher jerking levels (mean = 345 ms(-3) vs 172 ms(-3), p<0.05). Our newly implemented 3D approach is faster by 87.5% in extracting body motion trajectories when compared to a 2D frame by frame tracking procedure. We conclude that this new approach provides a more comfortable (both for patients and clinical professionals), simpler, faster and lower-cost procedure than previous approaches, therefore providing a reliable tool to quantitatively analyze MOI patterns of epileptic seizures in the routine of EMUs around the world. We hope this study encourages other EMUs to adopt similar approaches so that more quantitative information is used to improve epilepsy diagnosis.

A novel portable, low-cost kinect-based system for motion analysis in neurological diseases.

Many neurological diseases, such as Parkinson's disease and epilepsy, can significantly impair the motor function of the patients, often leading to a dramatic loss of their quality of life. Human motion analysis is regarded as fundamental towards an early diagnosis and enhanced follow-up in this type of diseases. In this contribution, we present NeuroKinect, a novel system designed for motion analysis in neurological diseases. This system includes an RGB-D camera (Microsoft Kinect) and two integrated software applications, KiT (KinecTracker) and KiMA (Kinect Motion Analyzer). The applications enable the preview, acquisition, review and management of data provided by the sensor, which are then used for motion analysis of relevant events. NeuroKinect is a portable, low-cost and markerless solution that is suitable for use in the clinical environment. Furthermore, it is able to provide quantitative support to the clinical assessment of different neurological diseases with movement impairments, as demonstrated by its usage in two different clinical routine scenarios: gait analysis in Parkinson's disease and seizure semiology analysis in epilepsy.