The nonlinear time-series analysis method, based on the recurrence plot theory, has received great attention from researchers and has been successfully used in multiple fields. However, traditional recurrence plots that use Heaviside step functions to determine the recursive behavior of a point in the phase space have two problems: (1) Heaviside step functions produce a rigid boundary, resulting in information loss; and (2) the selection of the critical distance, ε, is crucial; if the selection is inappropriate, it will result in a low-dimensional dynamics error, and as of now, there exists no unified method for selecting this parameter. With regard to the problems described above, the novelty of this article lies in the following: (1) when determining the state-phase point recursiveness, a Gaussian function is used to replace the Heaviside function, thereby solving the rigidity and binary value problems of the recursive analysis results caused by the Heaviside step function; and (2) texture analysis is performed on a recurrence plot, new ways of studying complex system dynamics features are proposed, and a system of complex system dynamic-like measurement methods is built.

Inspired by regression for detection task, We propose a new method called Pixel-Box. Prior work on regression for object detection can not match small pedestrian. Instead, Pixel-Box is only uses one feature map pixel of each channel to regress offset of boxes and then uses all feature map to predict which position has object. The model has two sibling network at the end of convolution output, but simply forward and backward once for both training and testing. It is an end to end system can directly optimized by loss function. To improve accuracy, we present a new backward method directly impact the model input and appropriate change the input image to re-forward the network. We evaluate this method on Caltech benchmark, also presenting competitive speed.