Dynamic Belief Fusion for Object Detection
This work addresses the challenge of improving object detection accuracy for computer vision applications by integrating multiple detectors, though it is incremental as it builds on existing fusion techniques.
The paper tackles the problem of fusing outputs from multiple heterogeneous object detectors by proposing Dynamic Belief Fusion (DBF), which dynamically assigns probabilities based on detector confidence and prior performance, resulting in significantly greater detection accuracy on ARL and PASCAL VOC 07 datasets compared to conventional methods and individual detectors.
A novel approach for the fusion of heterogeneous object detection methods is proposed. In order to effectively integrate the outputs of multiple detectors, the level of ambiguity in each individual detection score is estimated using the precision/recall relationship of the corresponding detector. The main contribution of the proposed work is a novel fusion method, called Dynamic Belief Fusion (DBF), which dynamically assigns probabilities to hypotheses (target, non-target, intermediate state (target or non-target)) based on confidence levels in the detection results conditioned on the prior performance of individual detectors. In DBF, a joint basic probability assignment, optimally fusing information from all detectors, is determined by the Dempster's combination rule, and is easily reduced to a single fused detection score. Experiments on ARL and PASCAL VOC 07 datasets demonstrate that the detection accuracy of DBF is considerably greater than conventional fusion approaches as well as individual detectors used for the fusion.