Real-Time ASoC: List of Objectives
I. A theoretical attention software framework that enables the fusion of multimodal information.
This first objective consists in developing a platform-independent model. This model is used to assess the required resources for the implementation of the attention model.
II. Design and implementation of the hardware/software co-processing system
This objective includes the implementation of a hardware reconfigurable architecture that is communicated with the PC for benchmarking purposes. The implementation of the torque-based attention model is its most relevant outcome.
III. Connection of the real-time adaptive hardware architecture with the linguistic resource
Once the torque-based attention system is having also access to the motion cues that provides the inferred model for the object of interest, this connection will provide values for motion, and others such as texture, color, shape, or size of the object.
IV. Real-time adaptive architecture for segmentation
This objective will aim at enabling real-time reliable computation for a segmentation algorithm.
V. Integrate the real-time visual attention and action
This final objective goal will be the integratation of the real-time system with the system developed including actions and motion cues, and any prior contextual knowledge.
This first objective consists in developing a platform-independent model. This model is used to assess the required resources for the implementation of the attention model.
II. Design and implementation of the hardware/software co-processing system
This objective includes the implementation of a hardware reconfigurable architecture that is communicated with the PC for benchmarking purposes. The implementation of the torque-based attention model is its most relevant outcome.
III. Connection of the real-time adaptive hardware architecture with the linguistic resource
Once the torque-based attention system is having also access to the motion cues that provides the inferred model for the object of interest, this connection will provide values for motion, and others such as texture, color, shape, or size of the object.
IV. Real-time adaptive architecture for segmentation
This objective will aim at enabling real-time reliable computation for a segmentation algorithm.
V. Integrate the real-time visual attention and action
This final objective goal will be the integratation of the real-time system with the system developed including actions and motion cues, and any prior contextual knowledge.