Abstract
The early diagnosis of cerebral palsy is an area which has recently seen significant multi-disciplinary research. Diagnostic tools such as the General Movements Assessment (GMA), have produced some very promising results, however these manual methods can be laborious. The prospect of automating these processes is seen as key in advancing this field of study. In our previous works, we examined the viability of using pose-based features extracted from RGB video sequences to undertake classification of infant body movements based upon the GMA. In this paper, we propose a new deep learning framework for this classification task. We also propose a visualization framework which identifies body-parts with the greatest contribution towards a classification decision. The inclusion of a visualization framework is an important step towards automation as it helps make the decisions made by the machine learning framework interpretable. We directly compare the proposed framework's classification with several other methods from the literature using two independent datasets. Our experimental results show that the proposed method performs more consistently and more robustly than our previous pose-based techniques as well as other features from related works in this setting. We also find that our visualization framework helps provide greater interpretability, enhancing the likelihood of the adoption of these technologies within the medical domain.
Highlights
T HE process of automating the recognition, analysis and reconstruction of complicated motion, such as human activity, has been an area of interest for researchers in many varied fields, due to its inherent ability to streamline intensive manual processes. [5]
In our previous works ( [30], [31]), we proposed different methods by which we could examine the feasibility of applying these technologies to the healthcare domain, to aid with the early prediction of cerebral palsy (CP) in infants
Some leading examples of how deep learning frameworks have been implemented for action recognition, human motion analysis and classification can be found in works exploring pose estimation [28] [12], [17], [21], [41] and part-based segmentation [17], [42], [44]
Summary
T HE process of automating the recognition, analysis and reconstruction of complicated motion, such as human activity, has been an area of interest for researchers in many varied fields, due to its inherent ability to streamline intensive manual processes. [5]. To make our proposed framework fully interpretable, an important aspect is the automatically generated visualization capable of relaying pertinent information to the assessor This visualization highlights the segmented body parts that are showing movement abnormalities, and are subsequently providing the most significant contribution towards the final classification result. Another issue found in the existing methods [2], [30], [31], [33], [40] is the lack of comparisons of performance between different approaches in the literature. This dataset reflects the intra-class variance and associated complexity found in carrying out manual assessments in a real-world clinical environment
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
