Table 4 SFSCF-Net ablation experiments on the Stanford Cars dataset.

From: Significant feature suppression and cross-feature fusion networks for fine-grained visual classification

Original image

OIG

Stage 1–3

Stage 4

Stage 5

SFSM

CFM

Accuracy rate (%)

√

–

–

–

–

–

–

92.7

–

√

√

√

√

–

–

93.0

√

√

√

√

–

–

–

92.9

√

√

√

√

–

√

–

93.2

√

√

√

√

–

√

√

93.5

√

√

√

√

√

–

–

93.4

√

√

√

√

√

√

–

94.2

√

√

√

√

√

√

√

94.5

  1. Significant values are in bold.
Back to article page