Ensemble-based Adaptive Single-shot Multi-box Detector
This work addresses object detection challenges for datasets with small and complex objects, such as construction site equipment, but is incremental as it builds directly on the existing SSD framework.
The paper tackles improving object detection accuracy in SSD by proposing adaptive default box selection based on dataset aspect ratios and an ensemble method using SSD components. Adaptive selection improves mean average precision by 3%, and ensemble-based SSD improves it by 8%, especially for small datasets and complex objects like construction equipment.
We propose two improvements to the SSD---single shot multibox detector. First, we propose an adaptive approach for default box selection in SSD. This uses data to reduce the uncertainty in the selection of best aspect ratios for the default boxes and improves performance of SSD for datasets containing small and complex objects (e.g., equipments at construction sites). We do so by finding the distribution of aspect ratios of the given training dataset, and then choosing representative values. Secondly, we propose an ensemble algorithm, using SSD as components, which improves the performance of SSD, especially for small amount of training datasets. Compared to the conventional SSD algorithm, adaptive box selection improves mean average precision by 3%, while ensemble-based SSD improves it by 8%.