Fingerspelling detection on ChicagoFSWild+ (test) benchmark leaderboard · SOTA2 Research

Fingerspelling detection on ChicagoFSWild+ (test)

100AP (IoU=0.1)GT

Evaluation Results

Method	Links
GT2021.04		100	100	100	74.4	70	50.5	63
GT2021.04		100	100	100	74.4	70	50.5	63
multi-task model2021.04		59	56.2	44.8	45	41.5	27.7	50.3
Our Model2021.04		59	56.2	44.8	45	41.5	27.7	50.3
BMN2021.04		58	54.9	43.7	43.3	40.1	26	47
BMN2021.04		58	54.9	43.7	43.3	40.1	26	47
Base 32021.04		56	52.5	42	42.6	39.6	26.4	47.7
Base 32021.04		56	52.5	42	42.6	39.6	26.4	47.7
Base 22021.04		44.3	36.6	24.7	32.3	26.5	15.2	39
Base 22021.04		44.3	36.6	24.7	32.3	26.5	15.2	39
Base 12021.04		27.8	8.2	2.5	21.1	9.3	2.9	26.7
Base 12021.04		27.8	8.2	2.5	21.1	9.3	2.9	26.7