Model-free 2D detection of unseen objects – HOPEv2

This leaderbord shows the ranking for Model-free 2D detection of unseen objects on HOPEv2. The reported time is the average time to generate predictions for all objects in a single image (averaged over the datasets).

Date (UTC) Submission Test image AP AP50 AP75 APS APM APL AR1 AR10 AR100 ARS ARM ARL Time (s)
2025-10-01 3PT-Detection (f.k.a IPT) RGB 0.568 0.715 0.661 0.100 0.203 0.632 0.570 0.644 0.644 0.100 0.215 0.719 39.255
2025-09-02 gfreedet2-2d_sam2.1-p RGB 0.452 0.657 0.487 0.044 0.101 0.519 0.480 0.536 0.536 0.044 0.122 0.615 2.039
2025-10-01 CNOS25 (YOLOE-DINOv3) RGB 0.452 0.707 0.488 0.100 0.092 0.522 0.475 0.513 0.513 0.100 0.116 0.591 0.103
2025-09-02 gfreedet2-2d_sam-p RGB 0.440 0.653 0.471 0.100 0.096 0.508 0.472 0.530 0.530 0.100 0.138 0.611 2.351
2025-10-01 NeMO-Detect (NeMO) RGB 0.411 0.563 0.466 0.022 0.090 0.468 0.439 0.456 0.456 0.022 0.105 0.520 8.315
2025-10-01 NOVASplat (FastSAM) (NOVASplatv_hope) RGB 0.399 0.631 0.418 0.089 0.058 0.465 0.442 0.494 0.494 0.089 0.112 0.568 0.335
2025-09-23 CTL (ctl-maskrcnn) RGB 0.385 0.573 0.434 0.089 0.064 0.447 0.406 0.440 0.440 0.089 0.063 0.513 0.534
2024-11-29 GFreeDet-SAM (hope_thre03) RGB 0.384 0.629 0.404 0.089 0.073 0.446 0.419 0.472 0.472 0.089 0.118 0.544 2.283
2025-09-01 gfreedet2-2d_fastsam RGB 0.369 0.561 0.392 0.089 0.034 0.426 0.409 0.470 0.470 0.089 0.090 0.538 0.287
2024-11-29 GFreeDet-FastSAM (fastsam_test) RGB 0.364 0.554 0.389 0.089 0.037 0.421 0.407 0.466 0.466 0.089 0.099 0.535 0.325
2025-08-24 CNOS (FastSAM) - Dynamic onboarding (Dynamic_onboarding) RGB 0.343 0.522 0.365 0.089 0.042 0.395 0.381 0.440 0.440 0.089 0.081 0.505 0.468
2025-08-24 CNOS (FastSAM) - Static onboarding (Static_onboarding) RGB 0.343 0.520 0.362 0.089 0.032 0.396 0.394 0.457 0.457 0.089 0.084 0.528 0.469

Tip: Hover over the numbers to see more decimal places.