Beijing Institute of Technology

Xiaopeng Chen

Qiang Huang

Xuechao Chen

Yang Xu

Zhangguo Yu

Weimin Zhang

Junyao Gao

Di Fan

Hyunwoo Kim

Siyuan Gou

The present application discloses a wide-field-of-view anti-shake high-dynamic bionic eye. A trajectory tracking method based on a bionic eye robot includes: establishing a linear model according to a bionic eye robot; establishing a full state feedback control system on the basis of the linear model; in the full state feedback control system, acquiring an angle and an angular acceleration required for a joint in a target tracking process of the bionic eye on the basis of a preset trajectory expectation value and a preset joint angle expectation value; the method further includes: adopting a linear quadratic regulator (LQR) to calculate a parameter K in the full state feedback control system, and minimizing energy consumption by establishing an energy function, so as to optimize the coordinated head-eye motion control of the linear bionic eye. The present application achieves the optimal control of the target tracking.

Wide-field-of-view anti-shake high-dynamic bionic eye

The present application provides a method and apparatus for pose planar constraining on the basis of planar feature extraction, wherein the method includes: inputting the acquired RGB color image and point cloud image into spatial transformation network to obtain two-dimensional and three-dimensional affine transformation matrixes; extracting the planar features of the transformed two-dimensional affine transformation matrix and three-dimensional affine transformation matrix; inputting the acquired planar features into the decoder and obtain the pixel classification of the planar features; clustering the vectors corresponding to the planar pixels to obtain the segmentation result of the planar sample; using planar fitted by the segmentation result to make planar constraint to the pose calculated by vision algorithm. The application combines RGB-D information to perform plane extraction, and designs a new spatial transformation network to transform two-dimensional color image and three-dimensional point cloud image.

Method and apparatus for pose planar constraining on the basis of planar feature extraction

The application discloses a multi-task learning incorporating dependencies method for bionic eye's face attribute recognition, which is as follows: Determine the first face attribute and the second face attribute for attribute recognition of facial image. Obtain the first recognition task branch and the second recognition task branch. Establish the task dependency between the first recognition task branch and the second recognition task branch to obtain the first transformed face attribute fully connected layer related to the second face attribute. and the second transformed face attribute fully connected layer related to the first face attribute. Feed the first transformed face attribute fully connected layer into the prediction layer to predict the first face attribute of facial image. And feed the second transformed face attribute fully connected layer into the prediction layer to predict the second face attribute of facial image. The multi-task learning incorporating dependencies method for bionic eye's face attribute recognition will be obtained according to the above steps.

Multi-task learning incorporating dependencies method for bionic eye's face attribute recognition

Growing community of inventors

Beijing, China

Xiaopeng Chen