VidAnomalyNet: An Efficient Anomaly Detection in Public Surveillance Videos Through Deep Learning Architectures
Bibliographic record
Abstract
In the contemporary era, computer vision applications assume significance due to their role in the real world.Video surveillance is one such application that has become indispensable with plenty of unprecedented applications.Detection of abnormal events from surveillance videos in real time has its importance in applications like traffic monitoring, crime investigation, public safety, healthcare and operations management to mention few.With the emergence of Artificial Intelligence (AI) automatic video surveillance is taken to the next level with sophistication in learning detection of anomalies.Particularly deep learning model like Convolutional Neural Network (CNN) is found more appropriate for image processing.However, as one size does not fit all, CNN does not provide acceptable accuracy unless it is enhanced with suitable number of layers and configurations.Towards this end, in this paper, we proposed a novel deep learning architecture known as VidAnomalyNet which is based on CNN model.It is designed to have more appropriate learning process and detection of anomalies from surveillance videos.We proposed a framework to exploit our VidAnomalyNet architecture for leveraging detection performance.We also proposed an algorithm known as VidAnomalyNet for Automatic Anomaly Detection (VAAD).Automatic anomaly detection in the context of video anomaly networks refers to the use of computational methods to automatically identify unusual or abnormal patterns within a sequence of video frames.The goal is to develop models that can distinguish between normal activities and unexpected events or anomalies.Video anomaly detection is crucial in various applications, including surveillance, industrial monitoring, and public safety.At present, this algorithm detects three classes of anomalies like fire, accident and robbery.It can be easily extended to identify more number of anomalies.We also explored MobileNetV1 with transfer learning by adding new layers to the base model for video anomaly detection.Our empirical study has revealed that VidAnomalyNet outperforms MobileNetV1.Highest accuracy achieved by the proposed model is 96.35%.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| 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 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
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".