Table 3. 1-NN classification accuracy in the visual object recognition study using the SURF features.

	Train on source No adapt.	Unsup. GFK [13]	DA Labels: S		Labels: S, T KEMA Kint	Train on target No adapt.
			OT-lab [15]	JDA [26]
lS		0	20	20	20
lT		0	0	$\widehat{y}$	3
C → A	21.4±3.7	35.3±3.2	43.5±2.1	40.7±4.0	47.1 ± 3.0	35.4±2.4
C → D	12.3±2.8	35.6±5.0	41.8±2.8	40.0±4.0	61.5 ± 2.8	65.1±1.9
A → C	35.3±0.5	32.9±2.5	35.2 ± 0.8	34.0±3.1	29.5±3.0	28.4±1.6
A → W	31.0±0.7	32.0±3.4	38.4±5.4	36.0±5.1	65.4 ± 2.7	63.5±2.6
W → C	21.7±0.4	27.7±2.4	35.5 ± 0.9	31.8±1.9	32.9±3.3	28.4±1.6
W → A	27.0±1.5	33.3±2.1	40.0±1.0	31.5±4.7	44.9 ± 4.5	35.4±2.4
D → A	19.0±2.2	33.0±1.3	34.9±1.3	32.9±2.9	44.2 ± 3.1	35.4±2.4
D → W	37.4±3.0	69.7±3.8	84.2 ± 1.0	80.0±4.1	64.1±2.9	63.5±2.6
Mean	22.3	37.4	44.2	40.9	48.7	44.4

C: Caltech, A: Amazon, D: DSLR, W: Webcam.

l_domain: number of labels per class.

$\widehat{y}$: predicted labels.