- Rating
- 3
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
TGTrack augments LoRAT-L with a temporal geometry branch that predicts target location, scale, and reliability from sparse trajectory histories. Geometry conditionally allocates an additional search crop, Gate–Hann adjusts its spatial prior, and raw visual confidence selects between regional hypotheses. Evaluation covers AntiUAV410 and direct transfer to AntiUAV.
Critical strengths
-
Geometry determines where to acquire additional evidence, while appearance determines the final hypothesis. Reliability-dependent Hann modulation is consistent with uncertainty about the predicted crop center.
-
TGTrack improves the matched LoRAT-L baseline by 2.86 AUC points, with a positive sequence-level bootstrap confidence interval. The 4.03-point improvement under direct cross-dataset transfer further supports practical value.
-
Constant-velocity and generic trajectory controls, memory sensitivity, gate diagnostics, and cross-region selection analysis help explain the gains. The reported computation accounts for the observed conditional Recrop frequency.
Critical weaknesses
-
The motion-model controls are useful, but different policies may trigger additional searches at different frequencies. Comparisons against simpler search policies under matched execution rates or computation are needed.
-
The proposed branch combines age encoding, confidence weighting, query pooling, horizon fusion, and motion correction. The generic Transformer comparison changes several elements together, leaving their individual necessity unclear.
-
The reported reliability analysis uses Stage-1 cached histories, whereas Gate–Hann matters specifically on executed Recrop frames. Its behavior on these harder, policy-dependent cases needs separate validation.
-
Evaluation uses two related benchmarks and one visual backbone. The tracker always outputs a box, while legacy precision rewards absent frames; although acknowledged, this limits conclusions about prolonged disappearance and recovery reliability.
Questions for authors
-
Do the gains persist against simpler additional-search policies under matched Recrop frequency or computational cost?
-
Which temporal-branch components contribute most, beyond the generic Transformer baseline?
-
How reliable is the gate specifically on executed Recrop frames and after long gaps without admitted observations?
-
Can you report visibility-conditioned results and validate transfer to another benchmark or visual backbone?
2 comments
Q3 (W3). Gate reliability during online Recrop and long observation gaps
We thank the reviewer for distinguishing cached-history analysis from online Recrop behavior. We analyze predicted gate values and center errors on valid online Recrop predictions, using the following gate intervals.
| Gate interval | Frames | Mean gate | Mean center error (px) |
|---|---|---|---|
| 734 | 0.3299 | 253.82 | |
| 1,005 | 0.5071 | 214.57 | |
| 1,442 | 0.7094 | 187.81 | |
| 3,868 | 0.8868 | 98.64 | |
| Overall | 7,049 | 0.7384 | 149.57 |
Higher gate intervals have lower mean center error, decreasing from 253.82 to 98.64 px. This provides online, group-level evidence that the gate carries information about localization accuracy.
Long observation gaps. We also compare mean gates with the GT-based reliability target used in training, grouping predictions by the elapsed time since the latest admitted trajectory state.
| Observation gap | Mean gate | Mean GT target |
|---|---|---|
| 2–5 frames | 0.658 | 0.537 |
| 6–15 frames | 0.653 | 0.553 |
| $\geq$16 frames | 0.565 | 0.616 |
Mean gate decreases from 0.658 to 0.565 as the gap increases, indicating more conservative spatial regularization after prolonged periods without admitted observations. As further quantified in the appendix, the predicted gate exhibits a strong Spearman correlation with localization accuracy (), supporting its role as a reliability indicator.
The learned gate serves as a relative reliability indicator for modulating the Recrop spatial prior, rather than an exact estimate of geometric prediction accuracy.
Q4 (W4). Visibility-conditioned evaluation and generalization
To assess visibility-conditioned performance and broader applicability, we provide additional evaluations on visible-target frames, two alternative tracking architectures, and zero-shot transfer to NfS-30.
Visibility-conditioned evaluation. We compare LoRAT-L and TGTrack on the full test set and on frames with valid visible targets.
| Subset | Base AUC | TG AUC | Base SA | TG SA |
|---|---|---|---|---|
| All frames | 63.45 | 66.31 | 64.56 | 67.47 |
| Visible (valid GT) | 64.39 | 67.41 | 65.52 | 68.58 |
Of 129,691 frames, 2,622 are target-absent (exist=0) and 41 have invalid GT boxes. Under this evaluation, they remain in the all-frame calculation with zero AUC and SA contributions. Metrics are averaged per sequence. On the remaining 127,028 valid-GT frames, TGTrack gains 3.02 AUC and 3.06 SA points. The gains therefore persist for visible-target localization.
Cross-framework evaluation. We evaluate integrations with OSTrack-256 and LETrack on AntiUAV410, comparing each augmented tracker with its own baseline.
| Tracking framework | Baseline AUC | Augmented AUC | AUC |
|---|---|---|---|
| LoRAT-L | 63.45 | 66.31 | +2.86 |
| OSTrack-256 | 56.07 | 57.21 | +1.14 |
| LETrack | 57.11 | 59.65 | +2.54 |
The gains of 1.14 AUC on OSTrack-256 and 2.54 AUC on LETrack support applicability beyond LoRAT-L.
Zero-shot cross-dataset transfer. On NfS-30, we directly reuse the AntiUAV410-trained model and inference parameters, without retraining or tuning.
| Method | AUC | Precision | NormP | OP50 | OP75 |
|---|---|---|---|---|---|
| LoRAT-L | 60.44 | 73.76 | 75.12 | 75.10 | 50.33 |
| TGTrack | 63.22 | 76.19 | 77.98 | 78.52 | 52.73 |
| +2.78 | +2.43 | +2.86 | +3.42 | +2.40 |
TGTrack gains 2.78 AUC points and improves all reported NfS-30 metrics. Together with the original +4.03 AUC transfer gain on AntiUAV, these results support applicability across the evaluated architectures and datasets.
We will include these results and clarifications in the revised manuscript. Many thanks for your constructive feedback.
Best regards,
On behalf of the authors
Q1 (W1). Comparison under comparable Recrop budgets
We thank the reviewer for recognizing the value of our decoupled search-allocation design. We address the four questions with additional budget-controlled comparisons, component ablations, online gate analysis, and visibility-conditioned and transfer evaluations.
To separate search-allocation quality from additional search frequency, we reevaluate the geometry strategies on AntiUAV410 under a common protocol. The learnable geometry models are independently retrained. All strategies use the same visual tracker, conditional Recrop mechanism, and fixed Hann weight (λh=0.45), without Gate–Hann. For the budget-controlled variants, eligible Recrop requests are periodically subsampled with a fixed per-sequence phase.
At 1.8762% Recrop, the proposed strategy gains 1.58 AUC points over Constant Velocity (1.8340%), with only 55 additional executions. It also gains 1.34 points over Generic Learnable Geometry while executing 280 fewer Recrops.
Even at a 0.9709% Recrop rate, it achieves 64.89 AUC, exceeding both alternatives. The results support an advantage in search allocation that is not solely attributable to more frequent additional visual searches.
Q2 (W2). Contributions of individual temporal components
The generic Transformer comparison changes several design choices together. We therefore evaluate six variants of full TGTrack, each removing one component, to examine their contributions separately.
Every removal lowers AUC in these runs. Learned Horizon Fusion has the largest effect (-1.95 points), followed by Age Embedding (-1.14) and Confidence Weighting (-0.98). Relative-Age Bias, Query Pooling, and Adaptive Motion Correction also contribute, with drops of 0.52, 0.94, and 0.79 points.
These results support explicit modeling of elapsed time, observation reliability, and historical motion beyond the generic trajectory Transformer.