RESUMO
Resampling detection plays an important role in identifying image tampering, such as image splicing. Currently, the resampling detection is still difficult in recompressed images, which are yielded by applying resampling followed by post-JPEG compression to primary JPEG images. Except for the scenario of low quality primary compression, it remains rather challenging due to the widespread use of middle/high quality compression in imaging devices. In this paper, we propose a new convolution neural network (CNN) method to learn the resampling trace features directly from the recompressed images. To this end, a noise extraction layer based on low-order high pass filters is deployed to yield the image residual domain, which is more beneficial to extract manipulation trace features. A dual-stream CNN is presented to capture the resampling trails along different directions, where the horizontal and vertical network streams are interleaved and concatenated. Lastly, the learned features are fed into Sigmoid/Softmax layer, which acts as a binary/multiple classifier for achieving the blind detection and parameter estimation of resampling, respectively. Extensive experimental results demonstrate that our proposed method could detect resampling effectively in recompressed images and outperform the state-of-the-art detectors.
RESUMO
OBJECTIVE: To study the safety and immunogenicity of the Bilive combined hepatitis A and B vaccine produced by Sinovac Biotech Co., Ltd. METHODS: Samples were selected from first year students of a senior high school (adults group) and first to fifth grade 1-5 students of 3 primary schools (children group). Those who were susceptible to both hepatitis A virus (HAV) and hepatitis B virus (HBV), HAV only or HBV only were assigned to group AB, A and B respectively and were vaccinated with three doses (0, 1 and 6 month schedule) of Bilive combined hepatitis A and B vaccine, inactivated hepatitis A vaccine and recombined hepatitis B vaccine respectively. The dosage for adult group was 500 U hepatitis A antigen and/or 10 micro g hepatitis B surface antigen and the dosage for children group was half the dosage of adult group. The potential adverse effects were observed within 72 hours after vaccination. Serum samples were collected for testing anti-HAV and anti-HBs at month 2 and 7 after the initial dose. RESULTS: The rates of local adverse effects were 0.58% and 2.56% in children AB group and adults AB group and the general adverse effects rates were 9.88% and 5.45% respectively. Both local and general adverse effect rates were not significantly different to the control group. The sero-conversion rate of anti-HAV in children and adults AB group reached 100%, one month after 3 doses. The geometric mean titer (GMTs) reached 33,910 mIU/ml and 23,435 mIU/ml respectively, significant higher than that in control group (group A). The sero-conversion rates of anti-HBs were 97.30% and 96.63%, and GMTs were 103 mIU/ml and 102 mIU/ml in children and adults AB group respectively. No significant difference on sero-conversion and GMT was observed when compared with control group. CONCLUSION: The Bilive combined hepatitis A and B vaccine had good safety profile, and the immunogenicity both on anti-HAV and anti-HBs was similar to that of separated components.
