Research Of Object Tracking Based On Sparse Hashing Algorithm

As more intelligent technology is demanded in daily life, visual tracking has been playing a more and more critical role in the computer vision field. It is used in various applications such as video surveillance, motion analysis, transportation navigation and human computer interface. Large numbers of effective tracking approaches have been proposed in the literature. Although the field has obtained good results under the efforts of many researchers, there are still a lot of problems to solve in target tracking task. So designing a tracking algorithm with high robustness is still challenging and worth researching.In this paper, we summarize the background, current researching status and future devel-opment, and then analyze the challenges in the target tracking task and introduce the typical tracking algorithms. Based on this, we propose two target tracking algorithms based on sparse hashing method.The first one is a novel tracking framework based on blocking and sparse hashing method. Different from the previous work, we treat object tracking as an Approximate Nearest Neighbor searching process in a binary space. Using the hash functions, the target templates and candi-dates can be projected into the Hamming space, facilitating the distance calculation and tracking efficiency. First, we integrate both the inter-class and intra-class information to train multiple hash function^ for better classification while most classifiers in previous tracking methods usu-ally neglect the inter-class correlation, which may cause the inaccuracy. Then, we introduce sparsity into the hash coefficient vectors for dynamic feature selection, which is crucial to select the discriminative and stable features to adapt to visual variations during the tracking process.The second one is a novel tracking framework based on discriminative and sparse hashing algorithm. To be specific, we make full use of the label information to assign a compact and dis-criminative binary code for each. And to deal with out-of-sample case, multiple hash functions are trained to describe the learned binary codes, and group sparsity is introduced to the hash projection matrix to select the representative and discriminative features dynamically.Extensive experiments on various challenging image sequences demonstrate the effective-ness and robustness of the two proposed trackers, and show that the proposed algorithm performs favorably against the state-of-the-art methods.