Intelligent Visual Tracking in Unstabilized Videos

Related Work

The algorithm proposed in the paper focuses on two major problems i.e. digital video stabilization and visual tracking. To the best of our knowledge, there is no work published for visual tracking in an unstabilized video. We have attempted this complex problem first time and solutions were provided for target detection and tracking in an unstabilized visual sequences. In (Verma et al., 2018) algorithm, we have developed an algorithm which stabilizes the input video using target templates and affine transformation. Once the frame is stabilized, L-1 minimization is computed under the frame work of sparse solution matrix. The energy of coefficients is used to solve the occlusion problem. In our next work (Verma et al., 2019), we proposed a combined algorithm to simultaneously stabilize the frames and track the target in unstabilized video after stabilizing using a single frame work of Speeded Up Robust Features (SURF) method. Now we have added Gabor Filter based auto-detection of target method in the proposed method for robustness of above algorithm. A dedicated hardware design was presented (Verma, Kamlesh, et al., 2019) for electronic image stabilization and visual tracking. A novel approach was developed in (Verma, Kamlesh, et al., 2018) for video stabilization using the detection of target. Camera intentional motion was preserved while rejecting unintentional camera motion in the algorithm developed in (Saxena, Harsh, et al., 2019) for video stabilization. Since besides these works, no other work has been published till date, we present the related work under two separate headings of video stabilization and visual tracking as below.