VIDEO MONITORING TIZIMLARIDA MAQSADLI KUZATUV JARAYONINING MATEMATIK MODELI VA OPTIMALLASHTIRILGAN ALGORITMLARI
Keywords:
video monitoring, maqsadli kuzatuv, ko'p kamerali tizim, MOTA, IDF1, DeepSORT, Hung algoritmi, kamera handoff, o'zaro korrelyatsiya, optimal topshiriq taqsimoti.Abstract
Ushbu maqolada video monitoring tizimlarida maqsadli kuzatuv jarayonining matematik modeli va optimallashtirilgan algoritmlari tadqiq etiladi. Tadqiqotning dolzarbligi shundaki, zamonaviy ko'p kamerali monitoring tizimlarida maqsadni uzluksiz kuzatish, bir kameradan boshqasiga o'tkazish (handoff) va resurslarni maqbul taqsimlash masalalari yetarli darajada hal etilmagan. Muammoni hal etish uchun ko'p kamerali tizim uchun o'zaro yopishishga asoslangan (correlation-based) kuzatuv matematik modeli, Hung algoritmi yordamida optimal topshiriq taqsimoti va DeepSORT-GP gibrid kuzatuv algoritmi taklif etiladi. MOT17, MOT20 va CamNeT ko'p kamerali benchmark to'plamlarida o'tkazilgan tajribalar MOTA = 78.6%, IDF1 = 81.3% va handoff muvaffaqiyat darajasi 94.2% natijalarini ko'rsatdi. Ilmiy yangilik — ko'p kamerali maqsadli kuzatuv uchun o'zaro korrelyatsiya matrisi asosidagi optimal topshiriq taqsimoti va Gaussian jarayon prognozi bilan boyitilgan DeepSORT gibrid algoritmini birlashtirishdan iborat.
References
[1] Yilmaz A., Javed O., Shah M. Object tracking: A survey // ACM Computing Surveys. — 2006. — Vol. 38, No. 4. — P. 13–45. DOI: 10.1145/1177352.1177355
[2] Ferryman J., Shahrokni A. PETS2009: Dataset and challenge // Proc. IEEE PETS. — 2009. — P. 1–6. DOI: 10.1109/PETS-WINTER.2009.5399556
[3] Luo W. et al. Multiple object tracking: A literature review // Artificial Intelligence. — 2021. — Vol. 293. — P. 103448. DOI: 10.1016/j.artint.2020.103448
[4] Ciaparrone G. et al. Deep learning in video multi-object tracking: A survey // Neurocomputing. — 2020. — Vol. 381. — P. 61–88. DOI: 10.1016/j.neucom.2019.11.023
[5] Wen L. et al. UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking // Computer Vision and Image Understanding. — 2020. — Vol. 193. — P. 102907. DOI: 10.1016/j.cviu.2020.102907
[6] Bewley A. et al. Simple online and realtime tracking // Proc. IEEE ICIP. — 2016. — P. 3464–3468. DOI: 10.1109/ICIP.2016.7533003
[7] Wojke N., Bewley A., Paulus D. Simple online and realtime tracking with a deep association metric // Proc. IEEE ICIP. — 2017. — P. 3645–3649. DOI: 10.1109/ICIP.2017.8296962
[8] Zhang Y. et al. ByteTrack: Multi-object tracking by associating every detection box // Proc. ECCV. — 2022. — P. 1–21. DOI: 10.1007/978-3-031-20047-2_1
[9] Du Y. et al. StrongSORT: Make DeepSORT great again // IEEE Trans. Multimedia. — 2023. — Vol. 25. — P. 8408–8419. DOI: 10.1109/TMM.2023.3240881
[10] Ristani E. et al. Performance measures and a data set for multi-target, multi-camera tracking // Proc. ECCV. — 2016. — P. 17–35. DOI: 10.1007/978-3-319-48881-3_2
[11] Fleuret F. et al. Multicamera people tracking with a probabilistic occupancy map // IEEE Trans. Pattern Anal. Mach. Intell. — 2008. — Vol. 30, No. 2. — P. 267–282. DOI: 10.1109/TPAMI.2007.1174
[12] Maksai A., Wang X., Fleuret F. Non-Markovian globally consistent multi-object tracking // Proc. IEEE ICCV. — 2017. — P. 2563–2573. DOI: 10.1109/ICCV.2017.278
