Robustness and Explainability of Image Classification Based on QCNN

Table 3

QCNN accuracy against different adversarial attacks.

Types	FGSM	CW	JSMA	DeepFool
Attacks	0.8906	0.8906	0.8438	0.8750
Bitplane1	0.7498	0.7465	0.7599	0.7364
Bitplane2	0.7236	0.7283	0.7681	0.7826
Bitplane3	0.7942	0.7563	0.8013	0.7169
Bitplane4	0.8045	0.7954	0.7842	0.7765
Bitplane5	0.9219	0.8112	0.8750	0.8906
Bitplane6	0.8906	0.9219	0.7188	0.9375
Bitplane7	0.9375	0.9062	0.8728	0.9062
Bitplane8	0.9531	0.7656	0.8594	0.9219