CNN with Embedding Transformers for Person Reidentification

Figure 2

Comparison diagram of cross-aliasing phenomenon in Market-1501. For each 1 × 1 × C vector (288 in total) in the feature map F (24 × 12 × C) of Figure 3, we calculate its cosine distance with ) among ; in order to compare with the PCB, we take n = 6, and cut F into six horizontally slice by using part-based global average pooling (PGAP). In a word, we calculate the cosine distance of each ) with the PGAP vector of each feature map, and display it to the color of the closest part. PCB, part-based convolutional baseline.