---
title: "Surveillance Video Annotation for Entrance Monitoring"
description: "Most surveillance frames carry nothing worth annotating. We filtered 60K down to 8K, then linked every detection to an employee ID."
url: "https://unidata.pro/cases/surveillance-video-annotation-for-entrance-monitoring/"
date_modified: "2026-06-01T18:00:26+03:00"
language: "en-US"
---
Task
----

A client needed to process surveillance footage from a factory entrance to enable automatic employee identification and matching with an access control system.

The dataset included video from three camera angles:

- two cameras inside the entrance area
- one monitoring the exit

### Goal:

Transform raw surveillance video into a structured dataset for:

- person detection
- identity matching (ID linkage)

### Key challenges:

- excessive volume of irrelevant frames
- inaccuracies in neural network pre-annotation
- need for precise alignment between visual data and employee IDs

Solution
--------

### 01. Video Preprocessing & Frame Reduction

Raw footage contained a large amount of non-informative data.

We introduced a filtering stage:

- removed up to 80% of irrelevant frames
- reduced dataset size from 50–60K to ~8K frames

This step increased efficiency and improved overall dataset quality.

### 02. Neural Pre-annotation with Manual Refinement

We combined automation with human validation:

- neural network used for initial person detection
- manual correction of false positives
- precise adjustment of bounding boxes

This hybrid approach balanced speed with accuracy.

### 03. Automated ID Matching Integration

To connect visual data with identity data, we:

- developed a script to match employee IDs
- aligned annotations with access control system records

This transformed the dataset from simple detection into a usable identification pipeline.

### **04. Validation & Quality Control**

A dedicated validation stage ensured consistency:

- verification of pre-annotation outputs
- correction of detection errors
- refinement of object boundaries

Special focus was placed on alignment between detected individuals and assigned IDs.

| Stage | Input | Workflow Scope | Main Quality Checks |
|---|---|---|---|
| Video Preprocessing | Raw surveillance footage | Frame filtering, data reduction | Relevance of frames / noise reduction |
| Frame Extraction | Filtered video | Selection of usable frames | Frame quality / coverage |
| Pre-annotation | Extracted frames | Neural network-based person detection | Detection accuracy / false positives |
| Manual Refinement | Pre-annotated data | Correction and bounding box adjustment | Boundary precision / consistency |
| ID Matching | Annotation + ID data | Automated linking of employees to detections | ID alignment accuracy |
| Validation & QA | Final dataset | Multi-stage verification and refinement | Consistency / identity matching quality |
| Final Delivery | Completed dataset | Packaging and integration readiness | System compatibility |

## Hero

**Industry and Use Case:** Surveillance & Security **Data:** 90 minutes of video from three cameras, approximately 50-60 thousand frames **Project Duration:** 2 week

## Main Title

Surveillance Video Annotation for Entrance Monitoring

## Description

We annotated 90 minutes of video footage from a factory entrance surveillance system, reducing the number of frames from 50-60 thousand to just 8 thousand. We implemented neural network-based pre-annotation, refined the data manually, and conducted final validation to ensure precise matching of employees with their IDs.

## Progress - Results - Quote

### Progress - Steps

**List of Steps:**

- **Number of days:** Video Preprocessing & Frame Reduction — **Step Description:** 2 days
- **Number of days:** Pre-annotation & Manual Refinement — **Step Description:** 5 days
- **Number of days:** ID Matching Integration — **Step Description:** 2 days
- **Number of days:** Validation & Final Delivery — **Step Description:** 3 days

### Results

**List of Results:**

- Frame volume reduced by ~80% (from 50–60K to ~8K)
- Faster annotation workflow due to pre-annotation
- Improved accuracy through filtering and manual refinement
- Reliable dataset for employee detection and ID matching

### Quote

**Quote:** In surveillance data, more frames don’t mean better results. The real impact comes from filtering noise, focusing on relevant moments, and ensuring every annotation aligns with identity data.

**Author:** Roman Lukoshin

**Position:** Speech and Generative Data Manager

[Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)