Room Security System Using Machine Learning with Face Recognition Verification
Bibliographic record
Abstract
Machine Learning (ML), an intelligent system known for its capacity to automate procedures by discerning patterns pertinent to specific tasks such as detection, prediction, and pattern recognition, is increasingly being used to advance biometric technologies. Among these, facial recognition, a subset of computer vision-based biometrics, is emerging as a robust security measure. The present study is centered on the design of a room security system that leverages facial recognition, rooted in a Convolutional Neural Network (CNN) architecture. The CNN model was constructed within the Tensorflow framework, employing the Keras library and Scikit-learn, all embedded within a Raspberry Pi system. The model was trained on 15 registered face classes, with an additional three unregistered classes used for biometric security testing. Performance was evaluated using the False Acceptance Rate (FAR) and False Rejection Rate (FRR), metrics that assess the system's ability to accurately verify authorized and unauthorized users. Findings demonstrated that the CNN model achieved a 97% accuracy rate in facial identification. Furthermore, biometric security testing of the CNN model using room security devices yielded optimal results at a threshold of 90%, with FAR=26.67%, FRR=9.33%, and an Equal Error Rate (EER) of 21.33%. It was observed that factors such as lighting, data variation, resolution, and positional changes during data sampling could impact the system's performance in realtime operations. It is therefore recommended that data collection and facial scanning be consistently conducted under identical environmental conditions to enhance the accuracy of the system. This study signifies a substantial stride in the development of advanced room security systems, thus contributing to the broader realm of secure access control systems.
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 | 0.001 |
| 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 | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
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".