• DR. P. V KUMAR
• A. SPANDANA
• J. SRAVANTHI
• M. SAI NAGENDRA

Worldwide information demonstrates that a noteworthy extent of rough fatalities stem from activity collisions. Reaction time for therapeutic help at mishap scenes intensely impacts the probability of survival and is to a great extent affected by human components. Given the predominance of video reconnaissance and cleverly activity frameworks, there is a developing require for robotized strategies to identify activity mischances, especially among computer vision analysts. By and by, Profound Learning (DL) strategies have illustrated impressive viability in complex visual errands, making them alluring for such applications. This think about points to create a DL-based computerized framework for identifying activity mishaps in video film. The proposed approach accept that activity mishap events can be characterized by visual highlights unfurling over time. Thus, the demonstrate engineering comprises of stages for extricating visual highlights and distinguishing transient designs. Visual and transient highlights are learned amid preparing utilizing convolutional and repetitive layers, leveraging both custom and freely accessible datasets. The strategy accomplishes a discovery precision of 98% on open activity mishap datasets, demonstrating strong execution independent of street setups.

COMPUTER VISION BASED CCTV ACCIDENT DETECTION USING DEEP LEARNING

"COMPUTER VISION BASED CCTV ACCIDENT DETECTION USING DEEP LEARNING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 5, page no.d114-d119, May-2024, Available :http://www.jetir.org/papers/JETIR2405311.pdf

"COMPUTER VISION BASED CCTV ACCIDENT DETECTION USING DEEP LEARNING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 5, page no. ppd114-d119, May-2024, Available at : http://www.jetir.org/papers/JETIR2405311.pdf