Method cwisardH+ on the SBM-RGBD dataset

Contact name: Maurizio Giordano

Contact email: maurizio.giordano@cnr.it

Contact Institution: high performance computing and networking - CNR

Method webpage: coming soon

Method processing time: ~8 fps (Train) ~2 fps (Classify) on a 4 GHz Intel Core i7 (OpenMP)

Method parameters: -k 0:0 -t 0.70:0.70 -z 64:128 -b 4:16 -O 1:1:1:1)

Reference: M. De Gregorio and M. Giordano, CwisarDH+: Background Detection in RGBD Videos by Learning of Weightless Neural Networks, in Battiato S., Farinella G., Leo M., Gallo G. (eds) New Trends in Image Analysis and Processing - ICIAP 2017. Lecture Notes in Computer Science, vol 10590, Springer, pp. 245--253, 2017.

Get all result files: here.

A) Sequence-by-sequence results

Name	Recall	Specificity	FPR	FNR	PWC	Precision	F-Measure
Bootstrapping
BootStrapping_ds	0.9426	0.9875	0.0125	0.0574	1.6659	0.8845	0.9126
adl24cam0	0.3310	0.9579	0.0421	0.6690	8.0174	0.3366	0.3338
bear_front	0.4344	0.9349	0.0651	0.5656	19.6902	0.7045	0.5374
fall01cam0	0.6510	0.9739	0.0261	0.3490	4.4015	0.5949	0.6217
fall20cam0	0.5044	0.9540	0.0460	0.4956	6.9155	0.3732	0.4290
ColorCamouflage
Cespatx_ds	0.9826	0.9992	0.0008	0.0174	0.1163	0.9615	0.9719
Hallway	0.8379	0.9962	0.0038	0.1621	1.2094	0.9235	0.8786
colorCam1	0.9938	0.9908	0.0092	0.0062	0.8509	0.9690	0.9812
colorCam2	0.9988	0.9535	0.0465	0.0012	2.5959	0.9469	0.9722
DepthCamouflage
DCamSeq1	0.7005	0.9973	0.0027	0.2995	1.9267	0.9389	0.8024
DCamSeq2	0.4279	0.9962	0.0038	0.5721	2.8621	0.8391	0.5668
Despatx_ds	0.9839	0.9942	0.0058	0.0161	0.6787	0.9449	0.9640
Wall	0.6162	0.9920	0.0080	0.3838	4.1520	0.8836	0.7261
IlluminationChanges
ChairBox	0.8831	0.9966	0.0034	0.1169	1.5766	0.9694	0.9242
Ls_ds	0.0000	0.9999	0.0001	0.0000	0.0108	0.0000	0.0000
TimeOfDay_ds	0.0000	0.9983	0.0017	0.0000	0.1718	0.0000	0.0000
genSeq1	0.9997	0.9710	0.0290	0.0003	2.5425	0.8323	0.9083
IntermittentMotion
Shelves	0.7475	0.9984	0.0016	0.2525	0.6503	0.9038	0.8182
Sleeping_ds	0.9307	0.9117	0.0883	0.0693	8.4413	0.7281	0.8170
abandoned1	0.9795	0.9823	0.0177	0.0205	1.7793	0.5865	0.7337
abandoned2	0.9988	0.9877	0.0123	0.0012	1.0981	0.9179	0.9566
movedBackground1	0.9314	0.9702	0.0298	0.0686	3.0378	0.3192	0.4754
movedBackground2	0.2640	0.8847	0.1153	0.7360	15.5038	0.1352	0.1789
OutOfRange
MultiPeople1	0.9206	0.9930	0.0070	0.0794	1.2660	0.9179	0.9192
MultiPeople2	0.9014	0.9964	0.0036	0.0986	1.1326	0.9572	0.9285
TopViewLab1	0.9462	0.9958	0.0042	0.0538	0.5833	0.8863	0.9153
TopViewLab2	0.9506	0.9956	0.0044	0.0494	0.6157	0.8958	0.9224
TopViewLab3	0.7608	0.9974	0.0026	0.2392	0.7681	0.8619	0.8082
Shadows
Shadows_ds	0.9795	0.9953	0.0047	0.0205	0.5702	0.9300	0.9541
fall01cam1	0.7870	0.9947	0.0053	0.2130	1.8362	0.9092	0.8437
genSeq2	0.9979	0.9869	0.0131	0.0021	1.1832	0.9056	0.9495
shadows1	0.9980	0.9830	0.0170	0.0020	1.4885	0.9055	0.9495
shadows2	0.9969	0.9786	0.0214	0.0031	1.8931	0.8806	0.9352

B) Average results for each category

Name	Recall	Specificity	FPR	FNR	PWC	Precision	F-Measure
Bootstrapping	0.5727	0.9616	0.0384	0.4273	8.1381	0.5787	0.5669
ColorCamouflage	0.9533	0.9849	0.0151	0.0467	1.1931	0.9502	0.9510
DepthCamouflage	0.6821	0.9949	0.0051	0.3179	2.4049	0.9016	0.7648
IlluminationChanges	0.4707	0.9914	0.0086	0.0293	1.0754	0.4504	0.4581
IntermittentMotion	0.8086	0.9558	0.0442	0.1914	5.0851	0.5984	0.6633
OutOfRange	0.8959	0.9956	0.0044	0.1041	0.8731	0.9038	0.8987
Shadows	0.9518	0.9877	0.0123	0.0482	1.3942	0.9062	0.9264
AVG.	0.7622	0.9817	0.0183	0.1664	2.8806	0.7556	0.7470