Share your thoughts, 1 month free Claude Pro on usSee more

AI-generated image detection on ConImageGen

99.6Average Accuracy

DTS-Det

Updated 18d ago

Evaluation Results

Method
DTS-Det 2026.07	99.6	100	100	100	100	100	100	99.6	99.6	100	100	100	100	100	100	100	100	99.2	99.2	99.5	99.5	96.8	96.9	100	100	99.5	99.5	99.5
DTS-Det 2026.07	94.2	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	93.3
SigLIP2+DINOv3 2026.07	89.1	96.5	97.2	98	97.9	92.7	92.7	90.4	90.5	97.1	97.1	97.4	97.4	98.4	98.4	98.2	98.2	63.4	49.2	88.6	86.3	65.5	53.4	91.3	91.3	81.1	78.7	85.4
AIDE 2026.07	87.3	97.6	97.6	85.4	86.4	76	76.4	89.9	89.8	97.7	97.7	97.6	97.6	94	93.9	97.8	97.9	82.7	80.2	77.4	77.2	50.4	43.5	97.2	97.3	91.1	90.9	86.6
SigLIP2+DINOv3 2026.07	84.5	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	82.4
DINOv3 2026.07	84.2	81.5	77.6	97.7	97.8	92.2	91.7	86.8	85.2	97.9	97.9	97.9	97.9	96.9	96.8	98.4	98.4	60.7	47.5	78.6	80.3	52.9	41.1	88	89.1	65.5	64.3	82
NPR 2026.07	79.7	93.3	93.5	78.8	79.5	81.4	82	84.2	83.5	81.7	82.3	83.3	83.2	83.5	83.3	83.9	84.1	59.6	57.8	88.4	88	64.3	62.1	87.5	87.6	65.6	65.4	79.4
F3Net 2026.07	78.9	83.7	83.3	85.9	86.3	83.6	84	90.2	90.1	92.9	93.1	90.7	90.9	70	69.8	91.4	91.2	55	52.8	78.2	78.4	52.1	43.7	87.5	86.6	64	60.6	78.2
UnivFD 2026.07	78.7	77.4	71.8	94.6	94.5	94.2	94.1	76.6	71.7	88.2	87.1	90.2	89.7	87.7	86.3	83.9	81.7	56.7	39.6	70.3	71.2	56	46.2	88.1	89.9	58.9	51.1	74.9
FreqNet 2026.07	76.9	93.3	93.5	76.6	78.6	76	77.5	82.3	82.6	89.6	89.7	89.8	89.3	77.5	78.6	85.6	85	64.6	63.2	72.2	71.2	55.7	55	72.9	71.4	63.1	62.2	76.8
NPR 2026.07	75.3	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	76.1
AIDE 2026.07	74.8	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	73.9
Xception 2026.07	70.7	66.3	66.9	73.1	72.6	73.8	74.2	80.2	80.5	86.2	85.7	84.7	84.8	65.6	64.9	81.2	80.9	58.9	57.8	63.2	62.7	56.3	55	68.6	67.8	67.8	65.7	70.7
C2P-CLIP 2026.07	70.4	71.6	65	91.1	91.4	90.4	90.8	74.8	71	74.4	68.9	76	72.8	76.7	75.4	74.4	69.4	53.7	61.9	57.7	65	52.4	56.4	65.2	73.1	56.7	66.8	71.4
UnivFD 2026.07	70.4	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	71.3
DINOv3-LoRA 2026.07	70.3	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	60.1
SigLIP2-LoRA 2026.07	68	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	64.5
ResNet50 2026.07	67.9	68.8	69.6	54.1	54.6	68.6	69.9	78	77.7	79.2	79.6	82.7	83	62.8	62	77.7	78.3	56	57	64.3	63.5	53.1	51.4	72.5	72.5	64.6	63	67.8
SigLIP2 2026.07	67.6	61	55.2	54.2	42	73.5	72.8	85.4	86.9	85.6	88.9	86.6	88.7	70.1	68.1	85.4	87.2	50.1	59.8	57.7	69.1	49.2	65.1	65.4	74	55.2	64.9	70.8
F3Net 2026.07	67.6	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	68.1
Xception 2026.07	66.3	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	66.4
C2P-CLIP 2026.07	64.1	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	63.8
ResNet50 2026.07	63.7	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	62.4
FreqNet 2026.07	63.2	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	61.3
Noiseprint++ 2026.07	50.2	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	67.9
Noiseprint++ 2026.07	49.7	48.9	63.9	54.6	69.4	53.7	68.4	53.4	68.6	48.3	63.6	47	62.4	48.9	64	48.9	63.8	52.7	67.4	51.2	67.1	50.9	66.9	39.3	53.8	48.1	62.1	64.7