IIPA-Net: Joint Illumination-Invariant and Pose-Aligned Feature Learning for Person Reidentification

Figure 8

Retrieved results of baseline and IIPA-Net on the low-light Market dataset. The first column images are the query. For each example, the upper row images are the results of baseline and the lower row images are of IIPA-Net. The images with a green border are the correct retrieved results, and those with a red border are the incorrect results.

(a)

(b)

(c)

(d)

(e)