Gait recognition which is one of the most important and effective biometric technologies has a significant advantage in long-distance recognition systems. For existing gait recognition methods, the template-based approaches may lose temporal information, while the sequence-based methods cannot fully exploit the temporal relations among the sequence. To address the above issues, we propose a novel multiple-temporal-scale gait recognition framework which integrates the temporal information in multiple temporal scales, making use of both the frame and interval fusion information. Moreover, the interval-level representation is realized by a local transformation module. Concretely, 3D convolution neural network (3D CNN) is applied in both the small and the large temporal scales to extract the spatial-temporal information. Moreover, a frame pooling method is developed to address the mismatch of the input of 3D network and video frames, and a novel 3D basic network block is designed to improve efficiency. Experiments demonstrate that the multiple-temporal-scale 3D CNN based gait recognition method can achieve better performance than most recent state-of-the-art methods in CASIA-B dataset. The proposed method obtains the rank-1 accuracy with 96.7% under normal condition, and outperforms other methods on average accuracy by at least 5.8% and 11.1%, respectively, in complex scenarios. CCS CONCEPTS • Security and privacy → Biometrics.

Conformations of (R)-alpha-(N6-adenyl)styrene oxide adducts at positions X6 in d(CGGACXAGAAG).d(CTTCTTGTCCG) and X7 in d(CGGACAXGAAG).d(CTTCTTGTCCG), incorporating codons 60, 61 (underlined), and 62 of the human n-ras protooncogene, were refined from 1H NMR data. These were the R(61,2) and R(61,3) adducts. Chemical shift perturbations were in the 5'-direction from the sites of adduction; large changes were observed for C5 H5 and H6 in the R(61,2) adduct. The styrene moieties were only partially defined by NOE data. Spectral overlap, particularly for the R(61,2) adduct, prevented complete assignments of the aromatic resonances; likewise, there were insufficient data to orient the CH2OH moieties. Ring flips were slow on the NMR time scale. For the R(61,2) adduct 260 restraints were obtained from NOE data at three mixing times using relaxation matrix analysis; for the R(61,3) adduct 230 restraints were obtained. Structures emergent from molecular dynamics/simulated annealing for the R(61,2) adduct converged to average and maximum pairwise rms differences of 1.3 and 1.7 A, respectively, while those for the R(61,3) adduct converged to average and maximum pairwise rms differences of 1.2 and 1.6 A. Sixth root residual indices of 7.5 and 6.8 x 10(-2) were measured between the refined structures and NOE intensities using relaxation matrix calculations for the R(61,2) and R(61,3) adducts, respectively. The styrene rings were in the 5'-direction from the lesion sites in the major groove. The preferred orientation calculated for the R(61,2) adduct placed the styrene ring edgewise and approximately orthogonal to C5, while that calculated for the R(61,3) duplex had the styrene ring approximately orthogonal to the major groove edges of base pairs A6.T17 and R-SOA7.T16.

No abstract

No abstract

The (S)-alpha-(N6-adenyl)styrene oxide adducts at positions X6 in d(CGGACXAGAAG). d(CTTCTTGTCCG) and X7 in d(CGGACAXGAAG).d(CTTCTTGTCCG), incorporating codons 60, 61 (underlined), and 62 of the human n-ras protooncogene, were examined by 1H NMR. These were the S(61,2) and S(61,3) adducts. Chemical shift perturbations were in the 3'-direction from the sites of adduction; upfield changes associated with the styrene aromatic ring current were noted for S-SOA6 H2 and H1', T16 N3H, H6, and CH3 resonances in the S(61,2) oligomer. In the S(61,3) oligomer, S-SOA7 H1', T16 H1', C15 N4Ha, and H5 shifted upfield. The styrene aromatic rings flipped rapidly on the NMR time scale; under these conditions the ortho and meta aromatic protons were equivalent. A sequence effect, in which the S(61,2) adduct equilibrated between two conformers, while the S(61,3) adduct exhibited only a single conformation, was observed. Potential energy minimization of the S(61,2) adduct major conformation yielded a structure in which the styrene ring was oriented in the 3'-direction and interacted primarily with the complementary strand. For the S(61,3) adduct, 291 restraints were obtained from NOE data at three mixing times using relaxation matrix analysis. The emergent structures refined to an average rms difference of 1.3 A, determined by pairwise analysis. These were compared to NOE intensity data; the calculated sixth root residual index was 9.2 x 10-2 at 250 ms. In the refined structure, the styrene ring was also oriented in the 3'-direction and interacted with the complementary strand. The minor conformation of the S(61,2) adduct was not identified. These results contrasted with the diastereomeric R(61,2) and R(61,3) adducts, which underwent slow ring flips on the NMR time scale and for which small sequence effects involving the minimum energy conformation of the styrene ring were observed.