Presenter(s)

Yunheng Liu, Jinnan Yan

Download

Download Project (654 KB)

Description

Object detection is crucial for real-world applications like the self-driving vehicle, search and rescue missions, and surveillance systems. Therefore, it is essential to accurately detect all objects in the field of view. While cutting-edge technologies like Mask R-CNN work in specific regions in images, therefore, some image regions are usually ignored one object is covered partially by the other. In our project, we improve the performance of object detection through a dual mechanism. In particular, our proposed framework removes the already-detected objects in the original image, then perform the detection process once again to force the attention to the ignorable regions. The final results are obtained by merging the two sets of detection results. We conduct experiments to demonstrate the effectiveness of the proposed framework.

Publication Date

4-24-2019

Project Designation

Honors Thesis

Primary Advisor

Van Tam Nguyen

Primary Advisor's Department

Computer Science

Keywords

Stander Symposium project

Improving Object Detection with Dual Mask R-CNN

Share

COinS