Roi-weighted primary frame filtering for efficient deep video surveillance

Begmatov, Shohruh, Arabboev, Mukhriddin, Nishanov, Akhram

Муҳандислик ва Иқтисодиёт · 2026-yil

Annotatsiya

Real-time video surveillance systems frequently rely on deep neural detectors and multi-object trackers.Although these models provide high detection and tracking quality, processing every frame is computationally expensive,especially for long-term multi-camera deployments. This paper presents an ROI-weighted primary filtering algorithm thatdecides whether an incoming video frame should be processed by a full deep model or skipped and propagated usingpreviously computed results. The method is motivated by the observation that ordinary mean absolute frame differencetreats all pixels equally. At the same time, surveillance decisions are usually more sensitive to changes in object regionsthan to background fluctuations. The proposed estimator constructs a spatial importance map from previously detectedbounding boxes, expands these regions by a dilation margin, and computes a normalized weighted frame-differencescore. A cost-minimization rule converts this score into a binary processing gate. Experiments on MOT20 sequencesshow that the ROI-weighted score amplifies object-related changes by 1.15x to 1.87x relative to global MAFD. Practicaldeployment scenarios demonstrate that more than 65% of frames can be handled by lightweight filtering, reducing computationalload by 56.69% and 61.31% in two surveillance cases. These findings indicate that object-aware frame filteringis a simple and effective pre-inference mechanism for resource-efficient video analytics.

Maqola ma’lumotlari
MualliflarBegmatov, Shohruh, Arabboev, Mukhriddin, Nishanov, Akhram
JurnalМуҳандислик ва Иқтисодиёт
NashriyotПлеханов номидаги Россия иқтисодиёти университетининг Тошкент шаҳридаги филиали
ISSN3060-463X, 3060-463X
Nashr sanasi2026-06-01
Jild4
Son6
TilIngliz
Indeksga qo‘shilgan2026-08-26

Maqolaning asl sahifasi

Kalit so‘zlar

video surveillance; frame skipping; ROI-weighted difference; object-aware filtering; multi-object tracking; edge AI; computational load reduction

Iqtibos keltirish

APABegmatov, Shohruh, Arabboev, Mukhriddin, & Nishanov, Akhram (2026). ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE. Муҳандислик ва Иқтисодиёт, 4(6). https://muhandislik-iqtisodiyot.uz/index.php/journal/article/view/3543
GOST (OAK)Begmatov, Shohruh, Arabboev, Mukhriddin, Nishanov, Akhram ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE // Муҳандислик ва Иқтисодиёт. – 2026. – T. 4, № 6.
MLABegmatov, Shohruh, Arabboev, Mukhriddin, and Nishanov, Akhram. “ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE.” Муҳандислик ва Иқтисодиёт, 4(6), 2026. https://muhandislik-iqtisodiyot.uz/index.php/journal/article/view/3543
ChicagoBegmatov, Shohruh, Arabboev, Mukhriddin, and Nishanov, Akhram. “ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE.” Муҳандислик ва Иқтисодиёт, 4(6) (2026). https://muhandislik-iqtisodiyot.uz/index.php/journal/article/view/3543
HarvardBegmatov, Shohruh, Arabboev, Mukhriddin, and Nishanov, Akhram (2026) ‘ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE’, Муҳандислик ва Иқтисодиёт, 4(6). https://muhandislik-iqtisodiyot.uz/index.php/journal/article/view/3543
BibTeX
@article{begmatovshohruh2026,
  title = {ROI-WEIGHTED PRIMARY FRAME FILTERING FOR EFFICIENT DEEP VIDEO SURVEILLANCE},
  author = {Begmatov, Shohruh and Arabboev, Mukhriddin and Nishanov, Akhram},
  journal = {Муҳандислик ва Иқтисодиёт},
  year = {2026},
  volume = {4},
  number = {6},
  url = {https://muhandislik-iqtisodiyot.uz/index.php/journal/article/view/3543}
}

Boshqa maqola uchun iqtibos tayyorlash

Муҳандислик ва Иқтисодиёт jurnalidan boshqa maqolalar

Муҳандислик ва Иқтисодиёт — barcha maqolalar · 2026-yil arxivi

Begmatov, Shohruh — boshqa maqolalari