MULTI-TASK VEHICLE DETECTION WITH REGION-OF-INTEREST VOTING

Abstract

Vehicle detection is a challenging problem in autonomous driving systems, due to its large structural and appearance variations. In this paper, we propose a novel vehicle detection scheme based on multi-task deep convolutional neural networks (CNNs) and region-of-interest (RoI) voting. In the design of CNN architecture, we enrich the supervised information with subcategory, region overlap, bounding-box regression, and category of each training RoI as a multi-task learning framework. This design allows the CNN model to share visual knowledge among different vehicle attributes simultaneously, and thus, detection robustness can be effectively improved. In addition, most existing methods consider each RoI independently, ignoring the clues from its neighboring RoIs. In our approach, we utilize the CNN model to predict the offset direction of each RoI boundary toward the corresponding ground truth. Then, each RoI can vote those suitable adjacent bounding boxes, which are consistent with this additional information. The voting results are combined with the score of each RoI itself to find a more accurate location from a large number of candidates. Experimental results on the real-world computer vision benchmarks KITTI and the PASCAL2007 vehicle data set show that our approach achieves superior performance in vehicle detection compared with other existing published works.

Existing System

Deformable partbased model (DPM), histogram of oriented gradients (HOG)

Proposed System

In this paper, we propose a novel vehicle detection scheme based on multi-task deep convolutional neural networks (CNN), region-of-interest (RoI) voting and multi-level localization, denoted by RV-CNN. Multi-task learning is designed to impose knowledge sharing while solving multiple correlated tasks simultaneously, boosting the performance of a part or even all of the tasks. In our method, the CNN model is trained on four tasks: category classification, bounding box regression, overlap prediction and subcategory classification. Here, we introduce the subcategory classification task to enforce the CNN model to learn a good representation for vehicles under different occlusions, truncations and viewpoints. We utilize the proposed concept of 3D Voxel Pattern (3DVP) in for subcategory classification. 3DVP is a kind of object representation that jointly captures key object properties which relates appearance, object pose, occlusion and truncation for rigid objects in the clustering process. Then each 3DVP is considered to be a subcategory.

Architecture

BLOCK DIAGRAM