Faster-RCNN with a compact CNN backbone for target detection in infrared images

Baussard, Alexandre; D’acremont, Antoine; Quin, Guillaume; Fablet, Ronan

doi:10.1117/12.2575756

Cited by 3 publications

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the importance of the security field is unquestionable and plays a major role in preventing accidents and disasters, maintaining social order, and protecting lives and property, and with the development of society, security is increasingly emphasized. Various security threats and criminal activities pose a threat to public safety and personal property, so the study of security technology and measures is an important way to meet the needs of society, so the field of security is worth being studied, the human body detection technology in the field of security will have a very good progress and the future of this paper, through the training of the model for the model and the deep learning, the use of Faster-RCNN algorithm is applied to detectron2 carrier, so that the human body detection is more accurate and has good results in the field of security [4].…”

Section: Introductionmentioning

confidence: 99%

Deploying human body detection technologies in security systems: An in-depth study of the FASTER-GCNN algorithm

Jiang

2024

ACE

View full text Add to dashboard Cite

The field of human body detection, a pivotal area in computer vision, merits comprehensive discussion. Remarkable advancements have been achieved in the techniques for human body detection over the past few decades, with significant applications spanning various sectors. This discussion delves into the potential of human detection technology within the realm of security - a field that necessitates efficient and accurate human detection technology to promptly identify potential threats, suspicious behaviors, or unusual activities. Deep learning-based human detection algorithms have substantially improved capabilities in this domain, facilitating real-time tracking and identification of the human form, thereby enabling security personnel to respond swiftly. This paper employs the Faster-RCNN algorithm for model training, utilizing the Information and Automation Research (INRIA) database. The deep learning-trained model proves highly accurate in human body detection, effectively recognizing human movements and behaviors. Such capabilities hold immense potential for implementation within the security sphere, including video surveillance systems and other similar applications where effectiveness is crucial.

show abstract

Section: Introductionmentioning

confidence: 99%