Hand Movement Analysis

Coordinate Based

Video Based

Unsupervised Learning


  • Here we can mainly address limitations of datasets

Clustering

Auto-encoders

Distinguish Typical and Atypical Behaviors

How many clusters does the frames of the video comes under?

What algorithm is best to cluster?

What feature extractors can be used to optimize the clustering?

Cluster concentration Analysis

Determine different types of atypical repetitive behaviors

What clustering algorithm is the best to identify type anomalous repetitive behavior?

What feature extractors can be used to optimize the clustering?

Cluster concentration Analysis

What features are prominent inside a limited dataset to identify the type of anomalous behavior?

Generating an equation for distinguishing typical and atypical repetitive actions

What features are prominent inside a limited dataset to derive the equation?

Analysing the sequence of the motor movement before justifying the repetitive action as atypical

Distinguish Typical and Atypical Behaviors

What feature extractors can be used to optimize the class determination?

How to make the algorithm unbounded with a large variance across different patients

a dual stream abnormal detection network ensembled by videos of motor movement and repetition detection modules

Unsupervised Learning

Clustering

Pattern Identification Typical and Atypical Behaviors

What feature extractors can be used to optimize the clustering?

How many clusters does the motor movement comes under?

What algorithm is best to cluster?

Cluster concentration Analysis

Detailed analysis of clustering features repetitive behaviors

What coordinate based features are prominent inside a limited dataset to identify the type of anomalous behavior?

What clustering algorithm is the best to identify type anomalous repetitive behavior?

What feature fusions are best to identifying the typical and atypical repetitive hand movements

Autoencoders

Generating an equation for distinguishing repetitive actions

What features are prominent inside a limited dataset to derive the equation?

Analysing the sequence of the motor movement before justifying the repetitive action as atypical

Distinguish Typical and Atypical Behaviors

What feature extractors can be used to optimize the class determination?

a dual stream abnormal detection network ensemble by novel pose trajectory and repetition detection modules

How to make the algorithm unbounded with a large variance across different patients