EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Robust Shot Detection Method for Regional Language Videos using Edge and Histogram Statistics

Avani Bhuva, Dhirendra Mishra

International Research Journal of Multidisciplinary Technovation · 22.05.2026 · 331-353

DOI: 10.54392/irjmt26320 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Research Journal of Multidisciplinary Technovation, online erschienen am 22.05.2026. Verfasst von Avani Bhuva und 1 weiteren Autor:innen. Die Arbeit nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The huge amount of digital content is being uploaded by users every day, and processing these videos specifically for indexing and retrieval purposes is extremely important. This research is contributing to society by providing a baseline approach for Gujarati news channel video retrieval for regional users. The Gujarati language is spoken by more than 22 million people worldwide. To retrieve these videos correctly, we need to process them frame by frame; hence, shot segmentation is an essential component, followed by keyframe extraction, and, lastly, extracting text features for the retrieval of news videos. A three approaches were developed to detect shot boundaries and identify frames containing meaningful Gujarati text in broadcast news video. The three methods were: (1) An Shot Boundary Detection (SBD) methods based on global threshold (SBD-GT) with 5 distinct frame level features (2) SBD methods employing statistical adaptive thresholding (SBD-AT) with 5 distinct frame level features (3) Development of the proposed Edge Histogram Adaptive Shot Detection (EHASD) system with the fusion of edge and histogram based features. EHASD applied Statistical Edge-Histogram Thresholding (SEHT) together to evaluate edge deviation and histogram difference to improve the detection rate of both abrupt and gradual transitions. The proposed research was evaluated using a TV9 Gujarati news video dataset with 45920 total frames from different categories like weather, Cricket, and politics. The proposed approach, EHASD, achieved a precision of 93.25% and a recall of 92.01%, surpassing the values obtained by the global and adaptive thresholds in SBD approaches. Additionally, the z test & paired t test was also performed on the mean and standard deviation to validate the results of the proposed approach. In addition, the research confirmed that the proposed method can be used further to perform key frame selection followed with text extraction for the video retrieval purpose.

Zitieren

APA

Bhuva, A., & Mishra, D. (2026). A Robust Shot Detection Method for Regional Language Videos using Edge and Histogram Statistics. International Research Journal of Multidisciplinary Technovation, 331-353. https://doi.org/10.54392/irjmt26320

Vancouver

Bhuva A, Mishra D. A Robust Shot Detection Method for Regional Language Videos using Edge and Histogram Statistics. International Research Journal of Multidisciplinary Technovation. 2026:331-353. doi:10.54392/irjmt26320

BibTeX
@article{Bhuva2026,
  author = {Bhuva, Avani and Mishra, Dhirendra},
  title = {A Robust Shot Detection Method for Regional Language Videos using Edge and Histogram Statistics},
  journal = {International Research Journal of Multidisciplinary Technovation},
  year = {2026},
  pages = {331--353},
  doi = {10.54392/irjmt26320}
}
Kostenlos starten →