Visual-attention gabor filter based online multi-armored target tracking

The multi-armored target tracking(MATT)plays a crucial role in coordinated tracking and strike.The occlusion and insertion among targets and target scale variation is the key problems in MATT.Most stateof-the-art multi-object tracking(MOT)works adopt the tracking-by-detection strategy,which rely on compute-intensive sliding window or anchoring scheme in detection module and neglect the target scale variation in tracking module.In this work,we proposed a more efficient and effective spatial-temporal attention scheme to track multi-armored target in the ground battlefield.By simulating the structure of the retina,a novel visual-attention Gabor filter branch is proposed to enhance detection.By introducing temporal information,some online learned target-specific Convolutional Neural Networks(CNNs)are adopted to address occlusion.More importantly,we built a MOT dataset for armored targets,called Armored Target Tracking dataset(ATTD),based on which several comparable experiments with state-ofthe-art methods are conducted.Experimental results show that the proposed method achieves outstanding tracking performance and meets the actual application requirements.

Visual-simulation region proposal and generative adversarial network based ground military target recognition

Ground military target recognition plays a crucial role in unmanned equipment and grasping the battlefield dynamics for military applications, but is disturbed by low-resolution and noisyrepresentation. In this paper, a recognition method, involving a novel visual attention mechanismbased Gabor region proposal sub-network(Gabor RPN) and improved refinement generative adversarial sub-network(GAN), is proposed. Novel central-peripheral rivalry 3D color Gabor filters are proposed to simulate retinal structures and taken as feature extraction convolutional kernels in low-level layer to improve the recognition accuracy and framework training efficiency in Gabor RPN. Improved refinement GAN is used to solve the problem of blurry target classification, involving a generator to directly generate large high-resolution images from small blurry ones and a discriminator to distinguish not only real images vs. fake images but also the class of targets. A special recognition dataset for ground military target, named Ground Military Target Dataset(GMTD), is constructed. Experiments performed on the GMTD dataset effectively demonstrate that our method can achieve better energy-saving and recognition results when low-resolution and noisy-representation targets are involved, thus ensuring this algorithm a good engineering application prospect.