Hierarchical attentive Siamese network for real-time visual tracking

One shot learning approach for cross spectrum periocular verification

The use of face mask during the COVID-19 pandemic has increased the popularity of the periocular biometrics in surveillance applications. Despite of the rapid advancements in this area, matching images over cross spectrum is still a challenging problem. Reason may be two-fold 1) variations in image illumination 2) small size of available data sets and/or class imbalance problem. This paper proposes Siamese architecture based convolutional neural networks which works on the concept of one-shot classification. In one shot classification, network requires a single training example from each class to train the complete model which may lead to reduce the need of large dataset as well as doesn't matter whether the dataset is imbalance. The proposed architectures comprise of identical subnetworks with shared weights whose performance is assessed on three publicly available databases namely IMP, UTIRIS and PolyU with four different loss functions namely Binary cross entropy loss, Hinge loss, contrastive loss and Triplet loss. In order to mitigate the inherent illumination variations of cross spectrum images CLAHE was used to preprocess images. Extensive experiments show that the proposed Siamese CNN model with triplet loss function outperforms the states of the art periocular verification methods for cross, mono and multi spectral periocular image matching.

Robust thermal infrared tracking via an adaptively multi-feature fusion model

When dealing with complex thermal infrared (TIR) tracking scenarios, the single category feature is not sufficient to portray the appearance of the target, which drastically affects the accuracy of the TIR target tracking method. In order to address these problems, we propose an adaptively multi-feature fusion model (AMFT) for the TIR tracking task. Specifically, our AMFT tracking method adaptively integrates hand-crafted features and deep convolutional neural network (CNN) features. In order to accurately locate the target position, it takes advantage of the complementarity between different features. Additionally, the model is updated using a simple but effective model update strategy to adapt to changes in the target during tracking. In addition, a simple but effective model update strategy is adopted to adapt the model to the changes of the target during the tracking process. We have shown through ablation studies that the adaptively multi-feature fusion model in our AMFT tracking method is very effective. Our AMFT tracker performs favorably on PTB-TIR and LSOTB-TIR benchmarks compared with state-of-the-art trackers.

Proposal-Based Visual Tracking Using Spatial Cascaded Transformed Region Proposal Network

Region proposal network (RPN) based trackers employ the classification and regression block to generate the proposals, the proposal that contains the highest similarity score is formulated to be the groundtruth candidate of next frame. However, region proposal network based trackers cannot make the best of the features from different convolutional layers, and the original loss function cannot alleviate the data imbalance issue of the training procedure. We propose the Spatial Cascaded Transformed RPN to combine the RPN and STN (spatial transformer network) together, in order to successfully obtain the proposals of high quality, which can simultaneously improves the robustness. The STN can transfer the spatial transformed features though different stages, which extends the spatial representation capability of such networks handling complex scenarios such as scale variation and affine transformation. We break the restriction though an easy samples penalization loss (shrinkage loss) instead of smooth L1 function. Moreover, we perform the multi-cue proposals re-ranking to guarantee the accuracy of the proposed tracker. We extensively prove the effectiveness of our proposed method on the ablation studies of the tracking datasets, which include OTB-2015 (Object Tracking Benchmark 2015), VOT-2018 (Visual Object Tracking 2018), LaSOT (Large Scale Single Object Tracking), TrackingNet (A Large-Scale Dataset and Benchmark for Object Tracking in the Wild) and UAV123 (UAV Tracking Dataset).