Corrections to “SiamFT: An RGB-Infrared Fusion Tracking Method via Fully Convolutional Siamese Networks”

oleh: Xingchen Zhang, Ping Ye, Shengyun Peng, Jun Liu, Ke Gong, Gang Xiao

Format: Article
Diterbitkan: IEEE 2019-01-01

Deskripsi

In <xref ref-type="bibr" rid="ref1">[1]</xref>, an error was created during the preparation of the final files. This only affected <xref rid="fig1" ref-type="fig">Figure 5</xref> and had no influence on the discussion and conclusions in the paper. The correct graphs of <xref rid="fig1" ref-type="fig">Figure 5</xref> are given below. Two examples of modality weights computed using the proposed method are presented. In the first one, the target (car) is clear in the first frame and gradually becomes unclear due to over-exposure. In the second example, it is difficult to see the target from visible images due to darkness, whereas it is easy to locate the target in infrared images. It can be seen clearly that the computed modality weights indeed indicate different reliability degrees of two modalities.