VOC2007 PRELIMINARY RESULTS

Key to Abbreviations

Abbreviation	Contact	Institution	Email/submission
Brookes	Lubor Ladicky	Oxford Brookes University	lladicky@brookes.ac.uk
Darmstadt	Mario Fritz	TU Darmstadt	fritz@mis.tu-darmstadt.de
INRIA_Larlus	Diane Larlus	INRIA Rhones-Alpes	diane.larlus-larrondo@inrialpes.fr
INRIA_Flat	Marcin Marszalek	INRIA Rhones-Alpes	marszale@inrialpes.fr/flat
INRIA_Genetic	Marcin Marszalek	INRIA Rhones-Alpes	marszale@inrialpes.fr/genetic
INRIA_Normal	Hedi Harzallah	INRIA Rhones-Alpes	hedi.harzallah@inrialpes.fr/normal
INRIA_PlusClass	Hedi Harzallah	INRIA Rhones-Alpes	hedi.harzallah@inrialpes.fr/plusclass
IRISA	Ivan Laptev	INRIA/INRIA Rennes	ilaptev@irisa.fr
Mannheim	Martin Bergtholdt	University of Mannheim	aragorn@uni-mannheim.de
MPI_BOW	Matthew Blaschko	MPI Tuebingen	matthew.blaschko@tuebingen.mpg.de/mpi_bow
MPI_Center	Matthew Blaschko	MPI Tuebingen	matthew.blaschko@tuebingen.mpg.de/mpi_center
MPI_ESSOL	Matthew Blaschko	MPI Tuebingen	matthew.blaschko@tuebingen.mpg.de/mpi_essol
Oxford	Ondrej Chum	University of Oxford	ondra@robots.ox.ac.uk
PRIPUVA	Julian Stoettinger	Vienna University of Technology/University of Amsterdam	julian@prip.tuwien.ac.at
QMUL_HSLS	Jianguo Zhang	Queen Mary University of London	jgzhang@dcs.qmul.ac.uk/HarLap
QMUL_LSPCH	Jianguo Zhang	Queen Mary University of London	jgzhang@dcs.qmul.ac.uk/LapSp
TKK	Ville Viitaniemi	Helsinki University of Technology	vviitani@cc.hut.fi
ToshCam_rdf	Jamie Shotton	Toshiba Corporate R&D Center, Japan/University of Cambridge	jamieshotton@gmail.com/RDF
ToshCam_svm	Jamie Shotton	Toshiba Corporate R&D Center, Japan/University of Cambridge	jamieshotton@gmail.com/SVM
Tsinghua	Dong Wang	Tsinghua University	wdong01@mails.tsinghua.edu.cn
UoCTTI	Pedro Felzenszwalb	University of Chicago/Toyota Technological Institute Chicago	pff@cs.uchicago.edu
UVA_Bigrams	Koen van de Sande	University of Amsterdam	ksande@science.uva.nl/UvA_Bigrams
UVA_FuseAll	Koen van de Sande	University of Amsterdam	ksande@science.uva.nl/UvA_FuseAll
UVA_MCIP	Koen van de Sande	University of Amsterdam	ksande@science.uva.nl/UvA_MixedColoredInterestPoints
UVA_SFS	Koen van de Sande	University of Amsterdam	ksande@science.uva.nl/UvA_SequentialForwardSelection
UVA_WGT	Koen van de Sande	University of Amsterdam	ksande@science.uva.nl/UvA_WeibullGaborTiles
XRCE	Florent Perronnin	Xerox Research Centre Europe	Florent.Perronnin@xrce.xerox.com

Classification Results: VOC2007 data

Competition "comp1" (train on VOC2007 data)

Average Precision (AP)

	aeroplane	bicycle	bird	boat	bottle	bus	car	cat	chair	cow	diningtable	dog	horse	motorbike	person	pottedplant	sheep	sofa	train	tvmonitor
INRIA_Larlus	0.626	0.540	0.328	0.475	0.178	0.464	0.696	0.442	0.446	0.260	0.381	0.340	0.660	0.551	0.772	0.131	0.291	0.367	0.627	0.433
INRIA_Flat	0.748	0.625	0.512	0.694	0.292	0.604	0.763	0.576	0.531	0.411	0.540	0.428	0.765	0.623	0.845	0.353	0.413	0.501	0.776	0.493
INRIA_Genetic	0.775	0.636	0.561	0.719	0.331	0.606	0.780	0.588	0.535	0.426	0.549	0.458	0.775	0.640	0.859	0.363	0.447	0.506	0.792	0.532
MPI_BOW	0.589	0.460	0.313	0.590	0.169	0.405	0.672	0.402	0.443	0.283	0.319	0.344	0.636	0.535	0.757	0.223	0.266	0.354	0.606	0.406
PRIPUVA	0.486	0.209	0.213	0.172	0.064	0.142	0.450	0.314	0.274	0.123	0.143	0.237	0.301	0.133	0.620	0.100	0.124	0.133	0.267	0.262
QMUL_HSLS	0.706	0.548	0.357	0.645	0.278	0.511	0.714	0.540	0.466	0.366	0.344	0.399	0.715	0.554	0.806	0.158	0.358	0.415	0.731	0.455
QMUL_LSPCH	0.716	0.550	0.411	0.655	0.272	0.511	0.722	0.551	0.474	0.359	0.374	0.415	0.715	0.579	0.808	0.156	0.333	0.419	0.765	0.459
TKK	0.714	0.517	0.485	0.634	0.273	0.499	0.701	0.512	0.517	0.323	0.463	0.415	0.726	0.602	0.822	0.317	0.301	0.392	0.711	0.410
ToshCam_rdf	0.599	0.368	0.299	0.400	0.236	0.333	0.602	0.330	0.410	0.178	0.332	0.337	0.639	0.531	0.779	0.290	0.273	0.312	0.501	0.376
ToshCam_svm	0.540	0.271	0.303	0.356	0.170	0.223	0.580	0.346	0.380	0.190	0.275	0.324	0.480	0.407	0.781	0.234	0.218	0.280	0.455	0.318
Tsinghua	0.629	0.424	0.339	0.497	0.237	0.407	0.620	0.352	0.427	0.210	0.389	0.347	0.650	0.481	0.769	0.169	0.308	0.328	0.589	0.331
UVA_Bigrams	0.612	0.332	0.294	0.450	0.165	0.376	0.546	0.313	0.399	0.172	0.314	0.306	0.616	0.424	0.746	0.145	0.209	0.235	0.499	0.300
UVA_FuseAll	0.671	0.481	0.433	0.581	0.199	0.463	0.618	0.419	0.484	0.278	0.419	0.385	0.698	0.514	0.794	0.325	0.319	0.360	0.662	0.403
UVA_MCIP	0.665	0.479	0.410	0.580	0.168	0.440	0.612	0.405	0.485	0.278	0.417	0.371	0.664	0.501	0.786	0.312	0.323	0.319	0.666	0.403
UVA_SFS	0.663	0.497	0.435	0.607	0.188	0.449	0.648	0.419	0.468	0.249	0.423	0.339	0.715	0.534	0.804	0.297	0.312	0.318	0.674	0.435
UVA_WGT	0.597	0.337	0.349	0.445	0.222	0.329	0.559	0.363	0.368	0.206	0.252	0.347	0.651	0.401	0.742	0.264	0.269	0.251	0.507	0.297
XRCE	0.723	0.575	0.532	0.689	0.285	0.575	0.754	0.503	0.522	0.390	0.468	0.453	0.757	0.585	0.840	0.326	0.397	0.509	0.751	0.495

Precision/Recall Curves

Detection Results: VOC2007 data

Competition "comp3" (train on VOC2007 data)

Average Precision (AP)

	aeroplane	bicycle	bird	boat	bottle	bus	car	cat	chair	cow	diningtable	dog	horse	motorbike	person	pottedplant	sheep	sofa	train	tvmonitor
Darmstadt	-	-	-	-	-	-	0.301	-	-	-	-	-	-	-	-	-	-	-	-	-
INRIA_Normal	0.092	0.246	0.012	0.002	0.068	0.197	0.265	0.018	0.097	0.039	0.017	0.016	0.225	0.153	0.121	0.093	0.002	0.102	0.157	0.242
INRIA_PlusClass	0.136	0.287	0.041	0.025	0.077	0.279	0.294	0.132	0.106	0.127	0.067	0.071	0.335	0.249	0.092	0.072	0.011	0.092	0.242	0.275
IRISA	-	0.281	-	-	-	-	0.318	0.026	0.097	0.119	-	-	0.289	0.227	0.221	-	0.175	-	-	0.253
MPI_Center	0.060	0.110	0.028	0.031	0.000	0.164	0.172	0.208	0.002	0.044	0.049	0.141	0.198	0.170	0.091	0.004	0.091	0.034	0.237	0.051
MPI_ESSOL	0.152	0.157	0.098	0.016	0.001	0.186	0.120	0.240	0.007	0.061	0.098	0.162	0.034	0.208	0.117	0.002	0.046	0.147	0.110	0.054
Oxford	0.262	0.409	-	-	-	0.393	0.432	-	-	-	-	-	-	0.375	-	-	-	-	0.334	-
TKK	0.186	0.078	0.043	0.072	0.002	0.116	0.184	0.050	0.028	0.100	0.086	0.126	0.186	0.135	0.061	0.019	0.036	0.058	0.067	0.090
UoCTTI	0.206	0.369	0.093	0.094	0.214	0.232	0.346	0.098	0.128	0.140	0.002	0.023	0.182	0.276	0.213	0.120	0.143	0.127	0.134	0.289

Precision/Recall Curves

Classification Results: VOC2006 data

Competition "comp1" (train on VOC2007 data)

Average Precision (AP)

	bicycle	bus	car	cat	cow	dog	horse	motorbike	person	sheep
INRIA_Flat	0.665	0.748	0.832	0.666	0.482	0.463	0.372	0.746	0.661	0.649
INRIA_Genetic	0.683	0.741	0.849	0.682	0.487	0.482	0.375	0.741	0.664	0.635
Tsinghua	0.514	0.554	0.759	0.531	0.329	0.332	0.268	0.593	0.579	0.571
UVA_Bigrams	0.304	0.543	0.704	0.484	0.275	0.222	0.116	0.553	0.481	0.452
UVA_FuseAll	0.410	0.739	0.677	0.544	0.365	0.353	0.126	0.610	0.535	0.609
UVA_MCIP	0.502	0.743	0.662	0.543	0.331	0.385	0.136	0.611	0.527	0.554
UVA_SFS	0.404	0.726	0.698	0.572	0.385	0.346	0.133	0.609	0.556	0.528
UVA_WGT	0.215	0.500	0.713	0.491	0.304	0.252	0.111	0.433	0.480	0.610
XRCE	0.722	0.721	0.843	0.644	0.467	0.379	0.438	0.693	0.681	0.681

Precision/Recall Curves

Detection Results: VOC2006 data

Competition "comp3" (train on VOC2007 data)

Average Precision (AP)

	bicycle	bus	car	cat	cow	dog	horse	motorbike	person	sheep
IRISA	0.352	-	0.482	0.094	0.209	-	0.183	0.333	0.211	0.262
Oxford	0.568	0.360	0.535	-	-	-	-	0.539	-	-
UoCTTI	0.562	0.236	0.555	0.103	0.212	0.099	0.173	0.439	0.262	0.221

Precision/Recall Curves

Segmentation Taster (VOC2007 data)

Competition "comp5" (train on VOC2007 data)

Accuracy (% of pixels)

- Entries in parentheses are synthesized from detection results.

	[mean]	background	aeroplane	bicycle	bird	boat	bottle	bus	car	cat	chair	cow	diningtable	dog	horse	motorbike	person	pottedplant	sheep	sofa	train	tvmonitor
Brookes	8.5	77.7	5.5	0.0	0.4	0.4	0.0	8.6	5.2	9.6	1.4	1.7	10.6	0.3	5.9	6.1	28.8	2.3	2.3	0.3	10.6	0.7
(INRIA_Normal)	7.7	2.8	1.2	1.8	8.3	1.5	52.4	0.3	12.4	5.3	3.7	0.0	18.3	4.1	0.1	0.1	3.6	28.7	0.3	6.1	0.4	10.6
(INRIA_PlusClass)	23.5	2.9	0.6	44.8	34.4	16.4	19.9	0.4	68.0	58.1	10.5	0.4	43.5	7.7	0.9	1.7	59.2	37.2	0.0	5.5	19.0	63.2
(MPI_Center)	17.5	56.6	11.8	31.2	0.0	11.0	0.0	0.1	35.7	51.4	7.3	20.3	0.0	1.0	2.3	2.6	60.0	0.0	2.6	5.0	43.4	25.3
(MPI_ESSOL)	27.8	2.6	29.7	30.8	9.5	41.4	6.7	8.0	72.9	55.7	37.1	11.1	19.4	2.2	14.9	23.8	66.8	25.9	8.6	3.2	58.1	55.1
TKK	30.4	22.9	18.8	20.7	5.2	16.1	3.1	1.2	78.3	1.1	2.5	0.8	23.4	69.4	44.4	42.1	0.0	64.7	30.2	34.6	89.3	70.6
(UoCTTI)	21.2	2.5	24.1	52.5	0.4	1.6	16.4	49.4	32.6	1.0	5.5	9.5	0.1	0.2	2.7	20.9	60.2	11.4	0.0	25.8	71.7	57.5