Datasets
Radiological Hand Pose Estimation (RHPE)
We collect the Radiological Hand Pose Estimation (RHPE) data from Colombian patients of the Fundación Santa Fe de Bogotá. The database comprises images of radiographs taken from left and right hands of both male and female patients between 0 and 240 months of age (0-20 years), with bone age annotations made by two expert radiologists for each patient. The dataset is composed of 6,279 images divided into 3 sets: 5,487 for training, 713 for validation and 79 for testing, maintaining the proportion of images used in each split of the RSNA dataset. A similar bone age distribution between the datasets suggests that they are compatible and can be used to study the influence of ethnicity on bone age assessment algorithms.
Radiological Society of North America (RSNA)
In 2017, the Radiological Society of North America (RSNA) created the Pediatric Bone Age Challenge. The data was provided by Stanford University, the University of Colorado and the University of California - Los Angeles. The database comprises images in png format of radiographs taken from the non-dominant hand of both male and female patients between 0 and 240 months of age (0-20 years), with bone age annotations made by trained radiologists. The dataset is composed of 14,236 images divided into 3 sets: 12,611 for training, 1,425 for validation and 200 for testing.
Publications
SIMBA: Specific Identity Markers for Bone Age Assessments
Bone Age Assessment (BAA) is a task performed by radiologists to diagnose abnormal growth in a child. In manual approaches, radiologists take into account different dentity markers when calculating bone age, i.e., chronological age and gender. However, the current automated Bone Age Assessment methods do not completely exploit the information present in the patient's metadata. With this lack of available methods as motivation, we present SIMBA: Specific Identity Markers for Bone Age Assessment. SIMBA is a novel approach for the task of BAA based on the use of identity markers.
For this purpose, we build upon the state-of-the-art model, fusing the information present in the identity markers with the visual features created from the original hand radiograph. We then use this robust representation to estimate the patient's relative bone age: the difference between chronological age and bone age. We validate SIMBA on the Radiological Hand Pose Estimation dataset and find that it outperforms previous state-of-the-art methods. SIMBA sets a trend of a new wave of Computer-aided Diagnosis methods that incorporate all of the data that is available regarding a patient. To promote further research in this area and ensure reproducibility we will provide the source code as well as the pre-trained models of SIMBA.
Hand Pose Estimation for Pediatric Bone Age Assessment
We present a new experimental framework for the task of Bone Age Assessment (BAA) based on a local analysis of anatomical Regions Of Interest (ROIs) of hand radiographs. For this purpose, we introduce the Radiological Hand Pose Estimation (RHPE) Dataset, composed of 6,288 hand radiographs from a population that is different from the currently available BAA datasets. We provide Bone Age groundtruths annotated by two expert radiologists as well as bounding boxes and keypoints denoting anatomical ROIs annotated by multiple trained subjects. In addition to RHPE, we provide bounding boxes and ROIs annotations for the publicly available BAA dataset by the Radiological Society of North America (RSNA). We propose a new experimental framework with hand detection and hand pose estimation as new tasks to extract local information for BAA methods. Thanks to its fine-grained and precisely localized annotations, our dataset will allow to exploit local information to push forward automated BAA algorithms. Additionally, we conduct experiments with state-of-the-art methods in each of the new tasks. Our proposed model, named BoNet, leverages local information and significantly outperforms state-of-the- art methods in BAA. We provide the RHPE dataset with the corresponding annotations, as well as the trained models, the source code for BoNet and the additional annotations created for the RSNA dataset.