摘要:
Detection of a facial attribute such as a smile or gender in a human face in an image is performed by embodiments of the present invention in a computationally efficient manner. First, a face in the image is detected to produce a facial image. Facial landmarks are detected in the facial image. The facial image is aligned and normalized based on the detected facial landmarks to produce a normalized facial image. Local features from selected local regions are extracted from the normalized facial image. A facial attribute is predicted in each selected local region by inputting each selected local feature into a weak classifier having a multi-layer perceptron (MLP) structure. Finally, output data is aggregated from each weak classifier component to generate all indication that the facial attribute is detected in the facial image.
摘要:
Various embodiments are presented herein that may render an image frame on an autostereoscopic 3D display. A computer platform including a processor circuit executing a rendering application may determine a current orientation of a virtual camera array within a three-dimensional (3D) scene and at least on additional 3D imaging parameter for the 3D scene. The rendering application, with the aid of a ray tracing engine, may also determine a depth range for the 3D scene. The ray tracing engine may then facilitate rendering of the image frame representative of the 3D scene using a ray tracing process.
摘要:
Systems and methods of detecting an object using motion estimation may include a processor and motion estimation and object detection logic coupled to the processor. The motion estimation and object detection logic may be configured to include logic to detect an object in a frame of a video based on motion estimation. The video may include a first frame and a second frame. The motion estimation may be performed on a region of the second frame using sum of absolute difference between the region of the second frame and a corresponding region of the first frame.
摘要:
Systems and methods for providing next generation television with content shifting and interactive selectability are described. In some examples, image content may be transferred from a television to smaller mobile computing device, and an example-based. visual search may be conducted on a selected portion of the content. Search results may then be provided to the mobile computing, device. In addition, avatar simulation may be undertaken.
摘要:
Systems and methods may provide for obtaining or implementing augmented reality information. A logic architecture may be employed to detect a low acceleration condition with respect to an image capture device. The logic architecture may select data from a video associated with the image capture device in response to the low acceleration condition. The logic architecture may also use the data to obtain augmented reality information for the video. Additionally, the logic architecture may modify the video with the augmented reality information, or may display the video with the augmented reality information.
摘要:
Systems and methods of detecting an object using motion estimation may include a processor and motion estimation and object detection logic coupled to the processor. The motion estimation and object detection logic may be configured to include logic to detect an object in a frame of a video based on motion estimation. The video may include a first frame and a second frame. The motion estimation may be performed on a region of the second frame using sum of absolute difference between the region of the second frame and a corresponding region of the first frame.
摘要:
Detection of a facial attribute such as a smile or gender in a human face in an image is performed by embodiments of the present invention in a computationally efficient manner. First, a face in the image is detected to produce a facial image. Facial landmarks are detected in the facial image. The facial image is aligned and normalized based on the detected facial landmarks to produce a normalized facial image. Local features from selected local regions are extracted from the normalized facial image. A facial attribute is predicted in each selected local region by inputting each selected local feature into a weak classifier having a multi-layer perceptron (MLP) structure. Finally, output data is aggregated from each weak classifier component to generate all indication that the facial attribute is detected in the facial image.
摘要:
An auto stereo display system for subway tunnel is disclosed. When a train is traveling through the tunnel at high speed, passengers can see stereo images outside the window by the system. The system is composed of LED light columns that vertically aligned at equal distance on the tunnel wall. Each LED light column is composed of multiple columns of LEDs and slit gratings. When the LEDs are changing with the changing instant speed of multi viewing angle image pixel array and the train, the passengers can see stereo image without glasses. The display system is specially applicable for business advertisement and nonprofit propagation in subway tunnel.
摘要:
Techniques are disclosed that involve face detection. For instance, face detection tasks may be decomposed into sets of one or more sub-tasks. In turn the sub-tasks of the sets may be allocated across multiple image frames. This allocation may be based on a multiple layer, quad-tree approach. In addition, face tracking tasks may be performed.