Estimating motion reliability to improve moving object detection in the H.264/AVC domain

Abstract

This paper presents a new algorithm for moving object detection in the H.264/AVC compressed domain which relies on motion vector information. In contrast to other motion vector-based algorithms, special attention is paid to noisy motion vectors as they highly decrease the performance of these algorithms. We propose to estimate the reliability of motion vectors by comparing them with projected motion vectors from surrounding frames. As such, noisy motion vectors are localized. By combining this information with the magnitude of motion vectors, foreground objects are distinguished. Experimental results demonstrate that our algorithm achieves significantly better segmentation quality compared to other motion vector-based approaches

Region-of-interest determination and bit-rate conversion for H.264 video transcoding

“…The decoded 49th frame in the Carphone sequence and the ROI detection result by [18] are shown in Figure 10a, b. Figure 10c demonstrates the ROI extraction without skin color data.…”

“…After motion trajec-tories are extracted, they use prediction residuals to improve the robustness of the proposed video activity detection strategy. De Bruyne et al [18] estimate the reliable motion information to enhance the ROI detection based on the forward/ backward projection of motion vectors (MVs) and similarity correlations. In addition, Poppe et al [19] present a fast moving object detection algorithm that relies on the structure of the encoded bit-stream at the syntax level.…”

Temporal Motion Vector Filter for Fast Object Detection on Compressed Video

“…The concept of motionvector projection used in this work is analogue to the temporal analysis from the STF [15]. But while in STF a current motion-vector is recursively projected to previous frames, in [17] the motion-vectors from previous and forward frames are projected to current frame. While in STF each mv has only one projection path, in [17] several mv can be projected to the same block, what results in true mv being averaged by noisy mv that points to the same blocks.…”

“…But while in STF a current motion-vector is recursively projected to previous frames, in [17] the motion-vectors from previous and forward frames are projected to current frame. While in STF each mv has only one projection path, in [17] several mv can be projected to the same block, what results in true mv being averaged by noisy mv that points to the same blocks. This difference results in better STF temporal analysis performance, specially in scenes with several noisy mv surrounding true mv.…”

“…In [17] is presented a method to estimate the reliability of motion-vectors compressed in H.264/AVC format. Each block motion-vector is compared with motion-vectors projected from previous and forward frames to determine their likelihood of representing a true object motion.…”