Downloads · 30 days
0
pantelisagg1/cctv-video-forensics-literature-review
cctv-video-forensics-literature-review is a machine learning model from pantelisagg1. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Two in-depth research analyses on state-of-the-art methodologies for detecting manipulations in CCTV surveillance footage.
Downloads · 30 days
0
Access
Public
Updated Apr 24, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
.md129 KB · 99%
From the Hugging Face model README
Two in-depth research analyses on state-of-the-art methodologies for detecting manipulations in CCTV surveillance footage.
File: CCTV_Video_Forensics_Literature_Review.md (~59 KB)
Comprehensive analysis of SOTA methods for detecting spatial and temporal manipulations in heavily compressed (H.264/H.265) CCTV videos.
Covers:
File: CCTV_Timestamp_Forensics_Literature_Review.md (~64 KB)
Structured analysis of methods for detecting timestamp manipulation and temporal metadata spoofing, distinguishing pixel-level from code-level forgery.
Covers:
A sophisticated video forger must manipulate four distinct layers simultaneously: pixels, container metadata, stream timing, and environmental consistency. Most forgeries fail because they address only 1-2 layers. The forensic analyst's advantage is multi-layer cross-validation.
| Analysis Layer | What It Detects | Hardest to Forge? |
|---|---|---|
| Stream Timing (PTS/DTS) | Splicing, frame deletion, timeline shifting | ✅ Very hard — deeply embedded in bitstream |
| Environmental (shadows, weather) | Wrong timestamp date/time | ✅ Impossible to forge — physical ground truth |
| Container/Hex Structure | Metadata editing, re-muxing, tool signatures | ⚠️ Moderate — sophisticated tools can clean |
| OSD Pixels | Burned-in timestamp alteration | ⚠️ Moderate — requires careful re-rendering |
70+ publications spanning 2005–2026, plus ISO standards and European forensic guidelines.
Research documents for educational and academic purposes.