Abstract

Indoor scene classification concerns a paramount task in computer vision: categorizing an indoor environment like a kitchen or office into predefined classes. This paper, in its application, uses a mixed model of RCNN and YOLOv11 to address the greatest challenges it poses: complex layouts and diversity in lighting and objects. techniques used would break scenes into smaller parts to ensure related semantic elements can be distinguished well for object detection and classification. The hybrid model combines the real-time detection feature of YOLOv11 with the precision of RCNN to improve system performance. It is optimized using tools such as Open-CV, TensorFlow, and Kera’s to be used in real-time applications, including object tracking, dynamic object monitoring, and security enhancement. Benchmark evaluations show large improvements in terms of accuracy, processing speed, and robustness compared to the traditional methods.

Authors

Showkat A. Dar, P. Rekha, Davud Fazil, K. Harshitha, M. Giresh, V. Likhitha
GITAM University, India

Keywords

Indoor Scene Classification, Hybrid Model, RCNN YOLOv11, Object Detection Real-Time Processing, Dynamic Object Tracking