Vehicle Attribute Recognition for Normal Targets and Small Targets Based on Multitask Cascaded Network

Abstract

The interference of the complex background and less information of the small targets are two major problems in vehicle attribute recognition. In this paper, two cascaded networks of vehicle attribute recognition are established to solve the two problems. For vehicle targets with normal size, the multitask cascaded convolution neural network MC-CNN-NT uses the improved Faster R-CNN as the location subnetwork. The vehicle targets in the complex background are extracted by the location subnetwork to the classification subnetwork CNN for the classification. The implementation of this task decomposition strategy effectively eliminates the interference of the complex background in target detection. For vehicle targets with small size, the multitask cascaded convolution neural network MC-CNN-ST applies the network compression strategy and the multilayer feature fusion strategy to extract the feature maps. These strategies enrich the location information and semantic information of the feature maps. In order to optimize the nonlinear mapping ability and the hard-to-detect samples mining ability of the networks, the activation function and the loss function in the two cascaded networks are improved. The experimental results show that MC-CNN-NT for the normal targets and MC-CNN-ST for the small targets achieve the state-of-the-art performance compared with other attribute recognition networks.

Multi-source target detection based on pixel-level fusion and decision-level fusion

“…In recent years, scholars have also tried to combine the pixel-level fusion, the feature-level fusion and the decision-level fusion for the target detection [3]- [7]. The combination of three fusion methods is applied to the classification of the land cover [8].…”

Life prediction of thermal aging of adhesive for retard‐bonded prestressed tendons

“…Liu et al 14 proposed an improved Arrhenius model that suggested the activation energy decreased at low‐temperature and verified its accuracy in predicting the lifespan of rubber materials. The BP neural network model based on test degradation data is suitable for solving nonlinear problems, which has appeared in life prediction of material of various fields 4,15,16 …”

“…The BP neural network model based on test degradation data is suitable for solving nonlinear problems, which has appeared in life prediction of material of various fields. 4,15,16 At present, there is no unified life prediction model for different polymer materials. The theoretical model and numerical fitting for life prediction are still in the exploratory stage of development.…”

Vehicle detection based on improved multitask cascaded convolutional neural network and mixed image enhancement

“…However, HybridNet did not classify the detection vehicles. Liu et al [23] adopted Faster R‐CNN as the basic network structure and designed a multitask cascaded CNN for vehicle detection. Gong et al [24] proposed an improved YOLOv3‐tiny for real‐time vehicle detection.…”