Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2021
- Goals
- Record
- 973f99f1-944c-47fd-8191-6561ea4f4dba
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted research addressing difficulties in accessing IP cameras from different manufacturers within an advanced surveillance system. The source explains that each IP camera has an individual manufacturing preset access address, which prevents users from automatically integrating IP cameras from other manufacturers. The research proposed selecting IP camera access links through an extraction approach using natural language processing. UDINUS researchers Noersasongko, Edi; Shidik, Guruh Fajar; Nugraha, Adhitya; Andono, Pulung Nurtantio; and Kusuma, Edi Jaya participated in the activity.
The method used TF-IDF as an NLP extraction technique and combined it with a statistical approach to select reliable candidate addresses from IP-camera address data provided by the ISPY dataset. The experiment applied several schemes based on the percentage of degradation of the original access address. The proposed results comprised 242 links at 25%, 288 links at 50%, 363 links at 75%, and 388 links at 100%. Evaluation using 10-fold validation in several simulations reported average accuracy in the range of around 89% to 97%. A brute-force complexity evaluation also indicated that the selected access address had lower complexity than the original address.
The activity is presented as UDINUS research focused on an NLP-based method for selecting and integrating access links in an IP-camera surveillance system.