IKKA: Inversion Classification via Critical Anomalies for Robust Visual Servoing

Darya Pavlenko

cs.LGApr 9, 2026

IKKA: Inversion Classification via Critical Anomalies for Robust Visual Servoing

Authors: Darya Pavlenko

Organizations: Clausthal-Zellerfeld, Germany

Abstract

We introduce IKKA (Inversion Classification via Critical Anomalies), a topologically motivated weighting framework for robust visual servoing under distribution shift. Unlike conventional outlier handling, IKKA treats maverick points as structurally informative observations: points where small perturbations can induce qualitatively different control responses or class assignments. The method combines local extremality, boundary transversality, and multi-scale persistence into a single anomaly weight, W(x) = E(x) x T(x) x M(x), which modulates control updates near ambiguous decision regions. We instantiate IKKA in a CPU-only embedded visual-servoing pipeline on Raspberry Pi 4 and evaluate it across 230 reproducible runs under nominal and stress conditions. In stress scenarios involving dim illumination and transient occlusion, IKKA reduces the 95th-percentile lateral error by 24% relative to a hybrid baseline (0.124 to 0.094) while increasing throughput from 20.0 to 24.8 Hz. Non-parametric analysis confirms a large effect size (Cliff's delta = 0.79).

Figures & tables

Tracker	FPS	CPU (%)	$P95\,\|e_{x}\|$ (nom.)	$P95\,\|e_{x}\|$ (stress)
HSV only	28.1	38	0.037	0.320
HSV + MOSSE	26.0	48	0.013	0.148
HSV + KCF	22.0	66	0.019	0.171
HSV + CSRT	18.0	78	0.014	0.139
Hybrid (M $\to$ C)	20.0	62	0.011	0.124
Hybrid + IKKA (ours)	24.8	45	0.010	0.094

Table 1: Tracker comparison across 230 runs. Nominal: no occlusion, standard lighting. Stress: dim illumination or

1

–

2

s occlusion. IKKA applied to Hybrid (MOSSE

\rightarrow

CSRT) configuration.

Configuration	Median recovery [s]	IQR [s]
HSV only	0.33	0.22
MOSSE	0.21	0.11
KCF	0.62	0.25
CSRT	0.95	0.58
Hybrid	0.53	0.58
Hybrid + IKKA	0.38	0.19

Table 2: Median recovery time (seconds) after target loss across 30 occlusion/obstacle runs.

Figure 1: Counterexample on a synthetic three-class dataset with a Wada-like boundary intersection at

(0,0)

. Left : SVM support vectors (crosses) concentrate along pairwise boundaries, far from the triple junction. Centre : IKKA anomaly weight

W(\mathbf{x})

is maximised near the origin. Right : IKKA top-5 maverick points (stars) cluster near the true intersection (diamond); SVM support vectors do not.

Variant	$P95\,\|e_{x}\|$ (stress)	FPS	Anomalous runs
Baseline (Hybrid, no IKKA)	0.250	20.0	4 / 30
$E$ only	0.210	21.5	3 / 30
$T$ only	0.195	22.0	2 / 30
$M$ only	0.205	21.2	3 / 30
$W=E\times T\times M$ (ours)	0.094	24.8	1 / 30

Table 3: Ablation study: individual IKKA components vs. full weight under stress conditions (

n=30

occlusion/obstacle runs).

Panel of Figure 2 — Figure 2: IKKA mechanism: three components combine multiplicatively into anomaly weight $W$ , which modulates the bounded IBVS yaw-rate command $\tau_{t}$ . High $W$ damps unsafe control escalation near ambiguous observability boundaries.

Explore similar work

Utilizing Inpainting for Keypoint Detection for Vision-Based Control of Robotic Manipulators
Apr 14, 2026Sreejani Chatterjee, Venkatesh Mullur, Abhinav Gandhi +1Vision-Based Robot Control Keypoint Detection
Robotics Engineering Program, Worcester Polytechnic Institute, MA, USA
DiffusionVS: A Generative Framework for Robust Visual Servoing Based on Diffusion Policy
Jun 17, 2026Hongkang Cui, Rui He, Haoyao ChenDiffusion Policy Robot Policy Generalization
ART-VS: Adaptive Resolution Tiling for Vision Transformer Visual Servoing
Jun 17, 2026Alessandro Scherl, Bernhard Neuberger, Simon Schwaiger +3Vision Transformer Vision-Based Robot Control
Department of Computer Technology, University of Alicante, Spain. · Department of Industrial Engineering, UAS Technikum Vienna, Austria. · Automation and Control Institute, TU Wien, Vienna, Austria. +2

Apr 14, 2026J/K move · Enter open · S save

Utilizing Inpainting for Keypoint Detection for Vision-Based Control of Robotic Manipulators

Sreejani Chatterjee, Venkatesh Mullur, Abhinav Gandhi, Berk Calli

Robotics Engineering Program, Worcester Polytechnic Institute, MA, USA

Vision-Based Robot Control Keypoint Detection

Open paper AskPDF arXiv

We present a novel visual servoing framework for controlling a robotic manipulator in configuration space using only natural visual features. To train our data-driven keypoint detector, we attach ArUco markers along the robot body, use their centers as keypoint labels, and apply image inpainting to remove the markers and reconstruct the occluded regions. This produces automatically labeled, markerless robot images without requiring accurate camera calibration or robot models. At runtime, a second inpainting model reconstructs robot regions that are partially occluded, enabling continuous keypoint detection. An Unscented Kalman Filter (UKF) further improves temporal consistency and robustness of the keypoint estimates. We demonstrate successful model-free, vision-based control using natural robot features under both full visibility and partial occlusion. To show broader applicability, we also extend the perception pipeline to two-module and three-module soft origami arms and qualitatively evaluate keypoint detection and temporal tracking on these platforms.

Abstract

Explore similar work

Utilizing Inpainting for Keypoint Detection for Vision-Based Control of Robotic Manipulators

DiffusionVS: A Generative Framework for Robust Visual Servoing Based on Diffusion Policy

ART-VS: Adaptive Resolution Tiling for Vision Transformer Visual Servoing