Abstract:
Head pose estimation is crucial for gaze estimation and attention modeling. Head pose indicates drowsiness, distraction, and attention in drivers. It can also control in-car infotainment systems. In controlled environments, computer vision algorithms using RGB cameras can estimate head pose, but in a car, sudden illumination changes, occlusions, and large head rotations make it difficult. Depth cameras reduce these issues. Continuous head rotation trajectories depend on time. This observation inspired our temporal deep learning model for head pose estimation from point cloud. Point cloud data is used to extract discriminative feature representation from the face’s 3D spatial structure. BLSTM layers are added to frame-based representations. This model outperforms non-temporal algorithms using point cloud data and state-of-the-art RGB image models on the newly collected multimodal driver monitoring (MDM) dataset. Quantitatively and qualitatively, temporal information improves prediction accuracy and smoothness.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
