A Bone Age Assessment based on a hybrid Knowledge Distillation Paradigm
using single ROI
Abstract
Bone age assessment (BAA) has been widely used in clinical area from
pediatric endocrine diagnosis to estimating height of children for
adulthood. Advancements in technology has accelerated the evolution of
BAA methodologies, one of which is deep learning algorithms, which
overcome the drawbacks of conventional approaches. In spite of excellent
effectiveness of deep neural networks in detection of the correct class
for bone age, they have a significant degree of complexity due to the
numerous parameters they employ for each ROI. Therefore, they are not
suitable for implementation in edge devices with limited resources. In
this paper, we propose a BAA method using a hybrid knowledge
distillation (KD) paradigm in order to conquer this difficulty by
mapping different ROIs into a single ROI. In this regard, the student
receives knowledge from a teacher network that has been pre-trained on
six ROIs including bones of five fingers and the wrist, transfers the
knowledge of its final response layer and internal layers to the
student. Then, six student models each of which is constructed based on
just one of these ROIs, while receiving the information of the teacher
model by matching the feature maps of intermediate layers and last
output of teacher with those of student in distillation modules.
Empirical results on Digital Hand Atlas (DHA) report that our student
model trained on one ROI obtains 95% accuracy on 19 classes of bone age
makes it appropriate for medical IoT deployment. Also, it has
competitive performance compared to the other state-of-the-art BAA
studies, and performance analysis of our KD with three base KD models
indicates superiority of our introduced method.