Projection-based perceptual video hashing
| dc.contributor.author | Sandeep R. | |
| dc.date.accessioned | 2020-08-03T10:57:00Z | |
| dc.date.accessioned | 2023-10-20T07:26:43Z | |
| dc.date.available | 2020-08-03T10:57:00Z | |
| dc.date.available | 2023-10-20T07:26:43Z | |
| dc.date.issued | 2019 | |
| dc.description | Supervisor: P. K. Bora | en_US |
| dc.description.abstract | A perceptual hash function for a video generates a fixed-length binary string called the perceptual hash on the basis of the perceptual contents of the video. This hash must be robust to the manipulations that preserve the perceptual contents of the video and fragile to the modifications that vary the perceptual contents of the video. Developing a perceptual video hashing method satisfying the conflicting properties is a challenge. This thesis generates the perceptual hash from the video based on the projection onto a sub-space by utilizing both the spatial and temporal properties of the video.This thesis first generates a perceptual hash from the video using the 3D-radial projection of the pixels and assesses the differentiating capabilities and the perceptual robustness of the hash generated. This method is the 3D extension of the well established 2D radial projection based image hashing. The pixel luminance values of randomly located sub-cubes are radially projected to calculate the variance of the pixels along the radial lines. The hash is obtained in two different ways. In Scheme-I, the variances along the radial lines are averaged, and then projected onto the discrete cosine transform (DCT) basis to form a compact hash vector. This hash vector is binarized using the median-based quantization. In Scheme-II, the variances along the lines of each sub-cube are projected on the 2D DCT basis, and then averaged to form a compact hash vector. This hash vector is also binarized using the median-based quantization. The performance of this algorithm is assessed using the receiver operating characteristic (ROC) curves. Simulation results indicate that the method performs well against most of the typical content-preserving distortions but poorly against the addition of noise. It also performs well against the malicious attacks | en_US |
| dc.identifier.other | ROLL NO.10610228 | |
| dc.identifier.uri | https://gyan.iitg.ac.in/handle/123456789/1461 | |
| dc.language.iso | en | en_US |
| dc.relation.ispartofseries | TH-2127; | |
| dc.subject | ELECTRONICS AND ELECTRICAL ENGINEERING | en_US |
| dc.title | Projection-based perceptual video hashing | en_US |
| dc.type | Thesis | en_US |
Files
License bundle
1 - 1 of 1