Akademik Veri Yönetim Sistemi
International Conferences on Science and Technology Engineering Sciences and Technology, Budva, Karadağ, 7 - 09 Eylül 2022, ss.60-71
In this work, a deep learning model, a multitask network is proposed to localize dynamic traffic objects such as cars, motorcycles, bicycles, buses, trucks, static traffic objects such as color classified traffic lights and traffic signs, pedestrians, segment drivable area and detect lane lines. The network design has a unified architecture, one shared encoder for feature extraction and three decoders for three tasks. The multi-network proposed is trained and tested with BDD100K dataset. Evaluation results show that the proposed method is the fastest multinetwork on the dataset with 47.62 FPS. Around multi-networks, proposed work has the second place on drivable area segmentation and lane line detection. Dynamic object localization performance of the network is state-of-theart with 40% performance increase compared to other models.