Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
The aim of this project is to successfully develop and industrialise the "Facial Movement 3D Dynamic Quantitative Measurement Device", which is a commercial device that can provide dynamic indicators of facial movement, and can practically solve the evaluation problems of facial paralysis for doctors and patients, and has important clinical value and social benefits.
Based on the current optimal high-precision face key point technology, the effective high-precision automatic face recognition model conforming to the Asian face is established through specialised training with targeted marker data, time series training optimisation based on optical flow, and local image tracking technology. The 3D reconstruction technique is then used to reconstruct the 3D spatial coordinates of the face keypoints, so as to realise the multi-camera 3D tracking of the face keypoints. The displacement of the key points is used to respond to the facial muscle activity characteristics of facial paralysis patients.
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| training session | A total of 150 facial expression videos of healthy volunteers and facial palsy patients were collected as a training set for the algorithm |
| |
| test set | A total of 50 cases of healthy volunteers and patients with facial palsy were used as the test set for the algorithm |
|
| Name | Type | Description | Arm Group Labels | Other Names |
|---|---|---|---|---|
| non-intervention | Other | non-intervention |
|
| Measure | Description | Time Frame |
|---|---|---|
| Facial feature point displacement distance | Displacement of corresponding facial feature points when volunteers perform specific actions such as specific expressions (raising eyebrow, closing eyes, shrugging nose, smiling and whistling) | 3 months |
| Facial Feature Point Velocity | Velocity of the corresponding facial feature points of volunteers when accomplishing specific actions such as specific expressions (raising eyebrows, closing eyes, shrugging nose, smiling, whistling) | 3 months |
Not provided
Not provided
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Based on previous face data collection often use larger public datasets, such as the AFLW dataset with 20,000 images. we need the same order of magnitude of local ethnographic data for migration learning. At the same time, the data in the dataset is improved according to the need to improve the accuracy of the labelled data at key points. Take about 100 effective frames for each data set. Need to complete about 200 people of various types of face data acquisition. The test group data is generally designed to be more than 20% of the training data, so the number of test group samples is designed to be 50 groups.
| Name | Role | Phone | Extension | |
|---|---|---|---|---|
| Guodong Feng | Contact | 15001275266 | 010 | fengguodong2013@163.com |
| Name | Affiliation | Role |
|---|---|---|
| Guodong Feng | Peking Union Medical College Hospital | Principal Investigator |
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| Peking Union Medical College Hospital | Recruiting | Beijing | Beijing Municipality | 100730 | China |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D005158 | Facial Paralysis |
| ID | Term |
|---|---|
| D009059 | Mouth Diseases |
| D009057 | Stomatognathic Diseases |
| D010243 | Paralysis |
| D009461 | Neurologic Manifestations |
Not provided
Not provided
Not provided
Not provided
Not provided
| D009422 |
| Nervous System Diseases |
| D012816 | Signs and Symptoms |
| D013568 | Pathological Conditions, Signs and Symptoms |