Small Object Detection using Context and Attention

There are many limitations applying object detection algorithm on various environments. Specifically, detecting small objects is still challenging because they have low-resolution and limited information. We propose an object detection method using context for improving accuracy of detecting small objects. The proposed method uses additional features from different layers as context by concatenating multi-scale features. We also propose object detection with attention mechanism which can focus on the object in image, and it can include contextual information from target layer. Experimental results shows that proposed method also has higher accuracy than conventional SSD on detecting small objects. Moreover, for 300 x300 input, we achieved 78.1% Mean Average Precision (mAP) on the PASCAL VOC2007 test set.

Automated summary

The proposed object detection method using context and attention mechanism achieves higher accuracy in detecting small objects compared to conventional methods.