Search

Lying-pose human detection is an active research field of computer vision in recent years. It has a good theoretical significance and furthermore many applications, such as victim detection or home service robot. But the study on lying-pose human detection in low-altitude overlooking images have many unsolved problems owing to multiple poses, arbitrary orientation, in-plane rotation, perspective distortion, and time-consuming. In this paper, the proposed framework of human lying-pose detection is optimization and machine learning algorithms inspired by processes of neurobiology suggest and human vision system to select possible object locations. First, the proposed model effectively utilizes binarized normed gradient features to obtain the objectness rapidly based on the vision saliency. Further, deep-learning techniques based on the convolution neural network are trained for learning rich feature hierarchies, in order to obtain the object of lying-pose human from objectness estimation, unlike the classical sliding-window algorithm. Eventually, employed pyramid mean-shift algorithm and rotation-angle recovery method to find position and direction of human lying-pose. The experimental results show that our method is rapid and efficient, and that it achieves state-of-the-art results with our XMULP dataset.

In Bag-of-Words BoW based image retrieval, soft assignment SA assigns R-nearest visual words to a feature, which significantly enhances the performance of image retrieval. However, it requires to calculate the weight of each visual word according to the distance from feature to visual word. This method sometimes loses its power when the codebook size is small, since a smaller codebook will cause larger quantization error and lead the distance to be more imprecise. In this paper, instead of depending on distance, we present a novel method to calculate the similarity between features by counting the number of identical visual words assigned to them. We describe how to create the inverted index, and weight the score between matched features. We evaluate the proposed Multi-Matched Similarity MMS method on Holidays and Ukbench datasets. Experimental results demonstrate that our method significantly improves the retrieval performance and outperforms the SA approach on both datasets.

A large number of computer vision applications rely on camera calibration. Camera self-calibration which only depends on the relationship between corresponding points of a pair of images draws much attention for its simplicity. Almost all the camera self-calibration methods rely on the solution of Kruppa equations which are difficult to be directly solved. The state-of-the-art self-calibration algorithms usually convert the solution of these equations to non-linear optimization problem, traditional optimization methods usually have the drawback of convergent to local extreme. Artificial immune system (AIS) has the ability to fast convergent to global extreme. To address this problem, we proposed an artificial immune system based method which can fast convergent to the global optimization solutions. We demonstrate the performance of the proposed method with synthetic and real data.