Foveated Vision

We present FoveaVLA, a Vision-Language-Action (VLA) model that replaces the wrist camera with learned foveation actions. Wrist cameras provide close-up detail but restrict the view to the end-effector's surroundings. External cameras see the whole scene, but the policy receives them at low resolution, so the fine detail needed for manipulation is lost. FoveaVLA recovers this detail by predicting, alongside each action chunk, the interaction region to observe at the next policy query. The policy then receives crops of these regions, taken from the original-resolution images, together with the full views. We build FoveaVLA on $π_{0.5}$ in two configurations, Single and Dual, using one or two external stereo cameras with full and foveal RGB and depth inputs, and post-train both on DROID with RoboInter interaction annotations. Across four real-world task suites, FoveaVLA-Single achieves 70.8% success with one external RGB-D view, and FoveaVLA-Dual achieves 77.1%, versus 52.1% for the wrist-equipped $π_{0.5}$. Both match the baseline on plain pick-and-place and improve success with distractors, cloth manipulation, and far-away objects.

Manipulation requires both broad scene awareness and detailed local feedback, yet conventional camera rigs provide them through separate front and wrist cameras. We present Fisheye-VLA, a visual interface that brings these capabilities together using a single passive fisheye. A global view preserves the workspace, while local perspective crops direct detail toward the interaction. The key design question is where this local visual budget should go. We answer it through a controlled re-rendering study, comparing alternative crop directions on the same recorded observations. The study finds that end-effector-centered views capture most of the estimated benefit of a much larger candidate pool, motivating a compact allocation around both hands. Our interface uses calibrated end-effector projection and motion lead to track the crops, while a shared ray encoding preserves their spatial meaning as they move. Integrated with a pretrained VLA, it achieves 84% and 82% success in the two expanded tabletop regions, where some target placements extend beyond the front-camera coverage, and supports shelf and conveyor manipulation. Ablations show that local crops and their viewing directions become more important in the larger workspace regions. The results demonstrate that a single fisheye can support these manipulation tasks without physical wrist cameras.

Fine-grained visual understanding depends on local detail, yet visual encoders face a trade-off between costly full-image high-resolution processing and compact global encoding that can weaken such evidence. Inspired by human active vision, we separate where to look from what to encode. We focus on the latter and introduce FAVE (Foveated Adaptive Visual Encoding), a lightweight variable-resolution ViT that encodes externally selected regions at high acuity while preserving native geometry. We first isolate this encoding problem using oracle ground-truth crops in a controlled small-object regime. On ImageNet objects with a native maximum side of 96 pixels, FAVE improves Top-1 by 9.4 points over a fixed-resolution ViT on the same crop window with 12.7 times lower FLOPs. Increasing global resolution or backbone capacity does not recover the same operating point. We then integrate FAVE as a complementary local branch in FastVLM. Its local tokens are combined with FastVLM's global visual tokens, while the original global pathway and language model remain frozen. With at most 16 additional local tokens, FAVE improves TextVQA by 1.60 points and achieves a 3.3 times controlled TTFT speedup over SmolVLM2-2.2B. On GQA attribute questions, it improves FastVLM-1.5B by 1.31 points, extending the benefit beyond text while narrowing the gap to FastVLM-7B. Together, these results show that selectively allocating high-acuity local capacity provides an efficient complement to broader global representations and model scaling for fine-grained understanding of small objects, text, and attributes.

Dense semantic segmentation allocates computational resources uniformly across the entire image, regardless of scene complexity or task relevance. Inspired by biological vision, we investigate whether semantic understanding can be achieved more efficiently through digital foveated perception. We introduce a lightweight active-vision pipeline that combines saliency-driven fixation selection, high-resolution foveal observations, low-resolution contextual information, semantic accumulation, and adaptive computation. Beyond conventional dense prediction metrics, we use object-level evaluation to measure semantic understanding under sparse observations. On ADE20K-Object, a single foveated observation achieves 95.9% of the baseline Top-1 accuracy and 96.9% of the baseline Top-3 accuracy while requiring only 4.7% of the computational cost. At the scene level, semantic accumulation recovers 90.6% of the baseline object recall while using 58.6% of the computation. These results suggest that substantial semantic understanding can emerge from sparse observations when computation is allocated selectively, highlighting active vision as an efficient alternative to uniform dense processing and motivating evaluation protocols beyond conventional pixel-wise segmentation metrics.

Human visual search is serial: the fovea must land on a candidate to confirm it, and those landings form a scanpath. Whether multimodal large language models (MLLMs), given the same foveated input, search as humans do bears on their use as models of human vision and on attention-alignment scores. We compare three general-purpose MLLMs with human eye-movement scanpaths on goal-directed search (COCO-Search18), driving each model fixation by fixation through an identical, human-matched foveated view and assessing it along three axes: the decision of target presence, the efficiency of reaching the target, and the gaze process itself. The axes dissociate. On the decision and on target acquisition the models match or exceed humans, detecting present targets near ceiling and reaching them on the first saccade more often than people do. The gaze process is not human. Under the human-matched condition, all three share one signature: low-entropy, large-amplitude, self-consistent scanpaths that agree with themselves far more closely than two humans agree with each other. That is consistent with a single-pass, non-serial architecture rather than a limit of acuity. Matched retinal input reproduces where humans look but not how the looking unfolds in time, and no degradation regime recovers human-like search at human-like success. The gap sits on a process axis that answer-alignment and saliency metrics do not measure. Because they miss it, such metrics cannot certify human-like vision, and zero-shot models suit outcome and spatial questions but not temporal, process-level ones.

Much of the recent progress in image and video recognition has come at the cost of memory: larger models, increased resolution, and longer temporal contexts. An inevitable component is the quadratic (or larger) growth of memory and compute based on image resolution, which is a property of the grid sampling used in convolutional networks and vision transformers. In this work we study residual networks whose convolutional blocks have logarithmic-square growth instead, enabling them to process very high-resolution video quickly. The key insight is to use a residual architecture's residual stream as a high-resolution buffer, to which convolutional blocks only read and write via log-polar image warp operations. Layers adaptively focus on different parts of each frame, with very high resolution only near the focus point. A complete high-resolution representation is built up in the residual stream, analogous to eye saccades creating a complete picture in biological vision, and a theoretical construction is presented that eliminates the quadratic dependency of the residual stream resolution. Experiments demonstrate that our proposed HiResNets learn to foveate around scenes similarly to human vision, and have superior performance in difficult egocentric video recognition tasks, especially egocentric video with small objects and fine-grained recognition.

Robotic foundation models still need task-specific fine-tuning before deployment, and the fine-tuned policies often break under modest changes in scene layout, lighting, or nearby distractors. We trace this brittleness to shortcut learning: fine-tuning supervises actions but not the visual evidence the policy uses, so the policy can settle on scene-level correlations that predict the demonstrations without causing success. We propose Artificial Foveated Perception (AFP), a lightweight, policy-agnostic module that takes the same vision and language inputs as existing Vision-Language-Action and World Action Model pipelines and predicts task-conditioned masks over the relevant objects, the robot, and other action-critical regions. During fine-tuning the masks serve as an auxiliary grounding signal that aligns the policy's visual attention with task-relevant regions; the policy architecture is unchanged, and at inference the policy runs on the original observation stream with no AFP call in the control loop. In simulation with four robotic foundation models and on a real robot with $π_{0.5}$, AFP improves generalization under environmental perturbations, reduces overfitting, and shortens fine-tuning. Ablations over mask quality and grounding-loss design show that these gains come from directing policy learning toward task-relevant visual evidence. Code, data, and videos are available at https://apollo-lab-yale.github.io/26-CoRL-AFP-website/.

The human visual system (HVS) employs foveated sampling and eye movements to achieve efficient perception, conserving both metabolic energy and computational resources. Drawing inspiration from this robustness and adaptability, we introduce the Foveated Dynamic Transformer (FDT), a foveation-guided dynamic token-selection architecture that integrates these mechanisms into a vision transformer framework. The FDT exhibits strong resilience to various types of noise and adversarial attacks, despite not being explicitly trained for such challenges. This inherent robustness is achieved through the use of fixation and foveation modules: the fixation module identifies fixation points to filter out irrelevant information, while the foveation module generates foveated embeddings with multi-scale information. At the 50% fixation-budget setting, FDT achieves higher accuracy than DeiT-S (81.9% vs. 80.9%) while reducing multiply-accumulate operations by 34.57%, highlighting one operating point on its accuracy-efficiency trade-off. These attributes position FDT as an HVS-inspired step toward artificial neural networks that combine adaptive computation with improved resilience.

When humans view scenes without a specific task (free-viewing), they initially direct their eye movements toward the scene center and then fixate on people, text, objects being gazed at or grasped, and semantically meaningful regions. What these signature fixation patterns reflect and whether they optimize an underlying perceptual task remain unknown. We show that a computational agent with simulated foveation, trained to optimize scene comprehension, exhibits emergent human fixation signature patterns. In contrast, versions of the agent trained to search or classify scenes, or equipped with peripheral vision that was better or worse than human vision, predicted human fixation patterns less accurately. Thus, human free-viewing fixation patterns may emerge as a functional byproduct of optimizing scene comprehension under the biological constraints of foveated vision.

Vision-language models benefit from high-resolution images, but the increase in visual-token count incurs high compute overhead. Humans resolve this tension via foveation: a coarse view guides "where to look", while selectively acquired high-acuity evidence refines "what to think". We introduce Foveated Reasoner, an autoregressive vision-language framework that unifies foveation and reasoning within a single decoding trajectory. Starting from a low-resolution view, the model triggers foveation only when needed, retrieves high-resolution evidence from selected regions, and injects it back into the same decoding trajectory. We train the method with a two-stage pipeline: coldstart supervision to bootstrap foveation behavior, followed by reinforcement learning to jointly improve evidence acquisition and task accuracy while discouraging trivial "see-everything" solutions. Experiments show that the method learns effective foveation policies and achieves stronger accuracy under tight visual-token budgets across multiple vision-language benchmarks.

Autonomous highway driving, especially for long-haul heavy trucks, requires detecting objects at long ranges beyond 500 meters to satisfy braking distance requirements at high speeds. At long distances, vehicles and other critical objects occupy only a few pixels in high-resolution images, causing state-of-the-art object detectors to fail. This challenge is compounded by the limited effective range of commercially available LiDAR sensors, which fall short of ultra-long range thresholds because of quadratic loss of resolution with distance, making image-based detection the most practically scalable solution given commercially available sensor constraints. We introduce Telescope, a two-stage detection model designed for ultra-long range autonomous driving. Alongside a powerful detection backbone, this model contains a novel re-sampling layer and image transformation to address the fundamental challenges of detecting small, distant objects. Telescope achieves $76\%$ relative improvement in mAP in ultra-long range detection compared to state-of-the-art methods (improving from an absolute mAP of 0.185 to 0.326 at distances beyond 250 meters), requires minimal computational overhead, and maintains strong performance across all detection ranges.

Semantics are one of the primary sources of top-down preattentive information. Modern deep object detectors excel at extracting such valuable semantic cues from complex visual scenes. However, the size of the visual input to be processed by these detectors can become a bottleneck, particularly in terms of time costs, affecting an artificial attention system's biological plausibility and real-time deployability. Inspired by classical exponential density roll-off topologies, we apply a new artificial foveation module to our novel attention prediction pipeline: the Semantic-based Bayesian Attention (SemBA) framework. We aim at reducing detection-related computational costs without compromising visual task accuracy, thereby making SemBA more biologically plausible. The proposed multi-scale pyramidal field-of-view retains maximum acuity at an innermost level, around a focal point, while gradually increasing distortion for outer levels to mimic peripheral uncertainty via downsampling. In this work we evaluate the performance of our novel Multi-Scale Fovea, incorporated into SemBA, on target-present visual search. We also compare it against other artificial foveal systems, and conduct ablation studies with different deep object detection models to assess the impact of the new topology in terms of computational costs. We experimentally demonstrate that including the new Multi-Scale Fovea module effectively reduces inherent processing costs while improving SemBA's scanpath prediction accuracy. Remarkably, we show that SemBA closely approximates human consistency while retaining the actual human fovea's proportions.

Vision-Language Models (VLMs) typically assume a uniform spatial fidelity across the entire field of view of visual inputs, dedicating equal precision to even the uninformative regions. By contrast, human vision is neither uniform nor static; it is adaptive, selective, and resource-efficient. In light of this, we present the first systematic analysis of bio-inspired visual representation methods, providing insights for more efficient and adaptive VLMs. We propose LLMind (Looking Like the Mind), a novel training-free framework that mimics foveated encoding and cortical magnification in human vision to achieve adaptive, efficient representations for VLMs under tight pixel budgets. Our key idea is to explore a Bio-inspired Adaptive Sampling Strategy (BASS), enabling a Mobius-parameterized module that performs non-uniform sampling while preserving global scene structure. On top of BASS, we introduce closed-loop semantic feedback (CSF) via test-time adaptation to align perceptual saliency with textual information from the frozen VLM. We evaluate LLMind against uniform and other sampling baselines across diverse scene-level and region-guided visual question answering benchmarks. The results show dramatic gains, with average improvements of +20% on VQAv2, +38% on Seed-Bench, and +37% on A-OKVQA compared to uniform sampling under tight pixel budgets. More surprisingly, LLMind retains up to 82%, 92%, and 97% of the full-resolution performance using only 1%, 3%, and 5% of the pixels, respectively. Moreover, LLMind is lightweight, plug-and-play, and compatible with existing VLMs without requiring architectural changes.

Human vision achieves remarkable perceptual performance while operating under strict metabolic constraints. A key ingredient is the selective attention mechanism, driven by rapid saccadic eye movements that constantly reposition the high-resolution fovea onto task-relevant locations, unlike conventional AI systems that process entire images with equal emphasis. Our work aims to draw inspiration from the human visual system to create smarter, more efficient image processing models. Using DINO, a self-supervised Vision Transformer that produces attention maps strikingly similar to human gaze patterns, we explore a saccade inspired method to focus the processing of information on key regions in visual space. To do so, we use the ImageNet dataset in a standard classification task and measure how each successive saccade affects the model's class scores. This selective-processing strategy preserves most of the full-image classification performance and can even outperform it in certain cases. By benchmarking against established saliency models built for human gaze prediction, we demonstrate that DINO provides superior fixation guidance for selecting informative regions. These findings highlight Vision Transformer attention as a promising basis for biologically inspired active vision and open new directions for efficient, neuromorphic visual processing.

Human eye movements in visual recognition reflect a balance between foveal sampling and peripheral context. Task-driven hard-attention models for vision are often evaluated by how well their scanpaths match human gaze. However, common scanpath metrics can be strongly confounded by dataset-specific center bias, especially on object-centric datasets. Using Gaze-CIFAR-10, we show that a trivial center-fixation baseline achieves surprisingly strong scanpath scores, approaching many learned policies. This makes standard metrics optimistic and blurs the distinction between genuine behavioral alignment and mere central tendency. We then analyze a hard-attention classifier under constrained vision by sweeping foveal patch size and peripheral context, revealing a peripheral sweet spot: only a narrow range of sensory constraints yields scanpaths that are simultaneously (i) above the center baseline after debiasing and (ii) temporally human-like in movement statistics. To address center bias, we propose GCS (Gaze Consistency Score), a center-debiased composite metric augmented with movement similarity. GCS uncovers a robust sweet spot at medium patch size with both foveal and peripheral vision, that is not obvious from raw scanpath metrics or accuracy alone, and also highlights a "shortcut regime" when the field-of-view becomes too large. We discuss implications for evaluating active perception on object-centric datasets and for designing gaze benchmarks that better separate behavioral alignment from center bias.

Human vision is a highly active process driven by gaze, which directs attention to task-relevant regions through foveation, dramatically reducing visual processing. In contrast, robot learning systems typically rely on passive, uniform processing of raw camera images. In this work, we explore how incorporating human-like active gaze into robotic policies can enhance efficiency and robustness. We develop GIAVA (Gaze Integrated Active-Vision ALOHA), a robot vision system that emulates human head and neck movement, and gaze adjustment for foveated processing. Extending the AV-ALOHA robot platform, we introduce a framework for simultaneously collecting eye-tracking, perspective control, and robot manipulation demonstration data from a human operator. We also open-source a simulation benchmark and dataset for training robot policies that incorporate human gaze. Inspired by recent work in foveated image segmentation and given the widespread use of Vision Transformers (ViTs) in robot learning, we integrate gaze information into ViTs using a foveated patch tokenization scheme. Compared to uniform patch tokenization, this significantly reduces the number of tokens, and thus computation. Our results show that our method for foveated robot vision drastically reduces computational overhead, and enhances robustness to background distractors. Notably, on certain high-precision tasks, foveated vision also improves performance, as reflected in higher success rates. Together, these findings suggest that human-inspired foveated visual processing offers untapped potential and should be further considered as a useful inductive bias in robotic vision systems. https://soltanilara.github.io/giava/