Accuracy			C.-Dec.		C.-Conf.
		QO	QS	QT	Scd	Rcd	Shape
Arch.	Model
CNN	ConvNeXt L	0.996	0.838	0.969	0.563	0.907	0.348
	RegNetY	0.996	0.797	0.986	0.546	0.895	0.456
	ResNeXt101	0.995	0.678	0.869	0.537	0.777	0.391
	DPN92	0.992	0.408	0.754	0.446	0.585	0.263
	ResNet101	0.989	0.326	0.848	0.363	0.593	0.181
	VGG19	0.983	0.228	0.774	0.304	0.510	0.124
Vision Transf.	EVA02 L	0.997	0.921	0.988	0.581	0.957	0.542
	BEiT	0.996	0.782	0.974	0.544	0.881	0.486
	ViT B16	0.994	0.699	0.931	0.528	0.820	0.455
	Swin B	0.994	0.683	0.911	0.527	0.802	0.321
	Inception v3	0.990	0.488	0.668	0.521	0.584	0.285
VLM	FLAVA-full	0.955	0.752	0.605	0.649	0.711	0.530
	Align-base	0.969	0.677	0.621	0.618	0.670	0.489
	CLIP ViT-B32	0.972	0.799	0.758	0.611	0.801	0.564
	SigLIP-base	0.975	0.829	0.816	0.602	0.843	0.478
	CLIP RN101	0.967	0.605	0.776	0.537	0.714	0.246
Hybrid	SEResNeXt	0.996	0.747	0.955	0.538	0.855	0.436
	CAFormer b36	0.997	0.758	0.983	0.534	0.873	0.415
	SENet154	0.992	0.427	0.812	0.439	0.625	0.261
	ConvFormer b36	0.994	0.477	0.955	0.426	0.720	0.277