Crime Scene Object Detection for Forensic Investigations Using Faster R-CNN and YOLOv5 Models
Bibliographic record
Abstract
More and more complex and numerous are the forensic findings required while investigating the crime scene, it is important to address the need for technological enhancements for the crime scene analysis. This objective detection models are important in helping alleviate the numerous errors that are involved when a human is tasked with the responsibility of identifying and categorizing objects within a crime scene, as it faster the process. In this research, Faster R-CNN and YOLOv5 deep learning models are used to detect the objects in crime scenes. Faster R-CNN which offers accuracy in object detection is used while YOLOv5 a real-time object detection framework improves the speed of detection. The models were trained and tested on a dataset which contains images of crime scene and the related objects include weapon, evidence mark and personal effect. The efficiency of the developed models was assessed by comparing the results on the mAP, detection speed, and false positive ratios. As such, the experimental results show that YOLOv5 is faster than Faster R-CNN for real-time applications, whereas Faster R-CNN is more accurate for higher detection rate-based applications. These models are complementary in their operation, the study suggests an integration of these models to improve efficiency for the forensic process. The study shows that by incorporating data object detection into contemporary forensic investigation processes powered by AI, the forensic science will significantly improve its ability to analyze evidence and solve crimes.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".