Overview

Image classification studies how to assign one or more labels to an image, supporting applications from species identification and medical diagnosis to document processing and remote sensing. Common approaches include convolutional neural networks, vision transformers, and pretrained vision-language models, trained with supervised, self-supervised, transfer, or few-shot learning. Researchers also develop data augmentation, hierarchical prediction, uncertainty estimation, and efficient models for settings such as ImageNet, corruption benchmarks, and specialized domain datasets. A central challenge is that strong accuracy on standard test sets may not transfer under distribution shifts or reveal reliance on spurious cues, while improving robustness, calibration, and efficiency can involve trade-offs.

Latest papers

422 papers35 papers in the last four weeks, down 8% on the four weeks before. 0.2% of all new papers.

Nov 13, 2026

Understanding Contextual Bias in Agricultural CNNs Through Hidden-Neuron Concept Induction

Convolutional Neural Networks (CNNs) can classify images using background-dominated patches that contain little or none of the target object, showing that they exploit contextual information. Previous studies detected this behavior through classification accuracy but say little about what CNN hidden layers represent. We investigate whether Concept Induction can reveal interpretable visual properties associated with hidden-neuron activations on contextual image patches. Using five agricultural datasets with high, moderate, and low contextual bias, we analyze four CNN architectures at three network depths and five patch locations. Quantitative image descriptors of each patch are encoded in a knowledge graph and used as background knowledge for Concept Induction, which derives logical expressions characterizing neuron activation. These expressions are tested on held-out images using the Mann-Whitney U test and Cliff's delta. Of 48,013 tested neuron-feature pairs, 13,164 (27.4%) were validated. Validation was most frequent at intermediate depths (mean 40.2% at mid layers versus 12.0% at deep layers), and no patch location dominated across datasets. Among the tested descriptors, edge, Gabor, and Tamura texture features had the highest validation rates. Stronger contextual bias did not correspond to higher neuron-feature validation rates: moderately biased Sugarcane had the highest rate (40.5%) and highly biased Cassava the lowest (19.2%). These results suggest that behavioral context dependence and the semantic interpretability of hidden neurons capture different aspects of CNN behavior, and that Concept Induction can complement patch-based bias detection.
Nov 13, 2026

SWAM: Saliency-Weighted Asymmetric Merging for Efficient Vision Transformers

Vision Transformers (ViTs) achieve strong performance on image recognition tasks, yet their quadratic self-attention complexity over input tokens creates prohibitive computational costs for practical deployment. Token reduction methods offer a practical solution by progressively reducing the token sequence during the forward pass. We propose Saliency-Weighted Asymmetric Merging (SWAM), a training-based token merging framework that extends the Bipartite Soft Matching (BSM) with two saliency-aware mechanisms. We introduce a two-step matching formulation in which cosine similarity decides candidate pairs, while a multiplicative maximum saliency penalty regulates their selection for merging. We then further use asymmetric exponential saliency-weighted aggregation, which allows more salient tokens to contribute more strongly to the merged representation. A lightweight MLP-based saliency prediction module (75k parameters per stage, 1% of backbone parameters) supervised via attention rollout is trained end-to-end jointly with a DeiT-Small backbone using a composite cross-entropy, knowledge-distillation, and saliency-alignment objective. On CIFAR-100, TinyImageNet, and ImageNet-100, under a conservative schedule that reduces tokens from 196 to 100 patch tokens, with the CLS token retained throughout, 27.5% FLOPs reduction and 1.35 times speedup, SWAM matches or exceeds baseline across the reported results. Under an aggressive schedule (196 to 70 tokens with the CLS token kept throughout, 33.8% FLOPs reduction, and 1.46 times speedup), accuracy gains are achieved on ImageNet-100 (+0.23%) with minimal losses on CIFAR-100 and TinyImageNet.
Nov 12, 2026

FLEX-LBP: Learning Local Binary Patterns with Exact Binary Inference for Texture Classification

Local Binary Patterns (LBP) remain among the most widely used texture descriptors, from industrial inspection to medical imaging. Yet their radius, sampling angles, color transform and thresholds are fixed by hand for each dataset. We present FLEX-LBP, which learns these parameters by backpropagation while every forward pass computes exact binary codes and hard histogram counts. A straight-through estimator acts only in the backward pass, so the deployed descriptor is an interpretable hard-binary LBP with 90 learnable parameters and unchanged inference cost. In a controlled ablation on six color texture benchmarks, the component worth learning is dataset-dependent. One configuration, applied unchanged everywhere, learns the geometry and thresholds with the color transform frozen. It improves the nearest-neighbor accuracy of the identical fixed descriptor by 3.55 points on average over five seeds, significantly on five of six benchmarks and never significantly worse. With a linear discriminant readout, it gains 0.9 points; the fully learned descriptor gains 2.7 and has the highest mean. FLEX-LBP is more accurate than five classical LBP-family descriptors, PCANet and a neural LBP on five of six benchmarks. Its mean nearest-neighbor accuracy exceeds that of frozen ImageNet ResNet-50 features (23.5 million parameters). On the Kather colorectal cancer histology benchmark, the fully learned descriptor reaches 93.8%, significantly above the best classical descriptor (Completed LBP, 92.8%). The learned radii adapt in directions that recur across seeds. Noise drowns single-pixel comparisons and illuminant change flips inter-channel ones; median sampling or gray-world normalization restores 96-97%. The code is available as a drop-in PyTorch layer at https://github.com/medalimoussa/FLEX-LBP.
Nov 10, 2026

picoEmergencyNet : Attention-Enhanced Ultra-Small Deep Learning Model for Aerial Image Classification in Emergency Monitoring

Real-time disaster recognition from aerial imagery requires deep learning models that operate under strict memory, compute, and latency constraints. While recent approaches achieve high accuracy, their computational cost limits deployment on resource-constrained UAV platforms. This work introduces picoEmergencyNet, an ultra-lightweight convolutional neural network designed for aerial image disaster classification. The proposed architecture combines depthwise dilated convolutions with efficient channel attention and channel shuffling to enable multi-scale feature extraction at minimal computational cost. picoEmergencyNet has 23k parameters and 7M FLOPs, requiring only 0.1 MB of model storage. Evaluated on the AIDER and AIDERv2 disaster datasets, it achieves up to 96.1% accuracy, matching or exceeding state-of-the-art lightweight models that are orders of magnitude larger. Deployment experiments on the NVIDIA Jetson AGX Orin show that, without hardware-specific optimizations, picoEmergencyNet achieves 258 FPS on GPU and 46 FPS on CPU with a 7x lower activation memory footprint compared to competing models. These results demonstrate that careful architectural specialization, rather than model scaling, enables highly efficient and accurate TinyML-ready disaster recognition for real-time UAV-based emergency monitoring.
Nov 9, 2026

GHiML: VLM-Guided Hierarchical Multi-Label Learning with Applications to Benthic Imagery

Vision language models (VLMs) encode substantial expert knowledge within their weights, yet their direct application to specialised classification domains such as benthic imagery remains limited. Meanwhile, vision transformers trained on these tasks rely predominantly on low-level visual latents (colours, textures, and shapes) that lack the higher-level reasoning humans employ for identification. We propose Guided Hierarchical Multi-Label (GHiML) learning, a pipeline that extracts conceptual latent variables from a VLM conditioned on training data, generates sample-relevant captions grounded in those variables, and fuses the resulting textual embeddings with image embeddings via cross-attention and a per-node gated router. We also introduce the inverted HML score (dinvd_{\mathrm{inv}}), a similarity score that evaluates hierarchical multi-label predictions through graph traversal distance and a cardinality penalty, providing an interpretable alternative to existing measures, which are insensitive to realistic degradations in sparse, deep hierarchies. Experiments on BenthicNet, FathomNet and iNaturalist show that the router improves macro F1F_1 over the vision-only baseline by up to 6.7 pp and dinvd_{\mathrm{inv}} by up to 21.1 pp on in-domain benthic data, with gains persisting under distribution shift. Ablations locate the contribution of the text itself: on the multi-label biotic hierarchies, removing the per-sample text signal degrades dinvd_{\mathrm{inv}} while leaving macro F1F_1 unchanged, consistent with text acting through prediction cardinality rather than node discrimination, whereas iNaturalist, whose annotations are single-label, loses 18.7% of macro F1F_1. Our results support the hypothesis that VLMs contain exploitable specialised knowledge that, when appropriately elicited, can augment classification in domains where direct visual learning is insufficient.
Nov 8, 2026

ADC-RANet: Improving Resolution Adaptive Neural Networks with Attention, Distillation and Calibration

Dynamic neural networks are a class of adaptive models in which the computation assigned to each input sample is determined dynamically, potentially reducing computational cost while maintaining similar performance. For image classification, the Resolution Adaptive Network (RANet) is a strong approach that progressively allocates computational resources across input scales based on sample difficulty. In this work, we introduce RANet with Attention, Distillation and Calibration (ADC-RANet), an enhanced version of RANet that improves performance across all computational budgets. Our approach combines three tightly connected components: bottom-up channel attention, top-down cascaded knowledge distillation, and overconfidence correction using MixUp, in a synergistic design that coordinates both pathways. Experiments on CIFAR-100 and ImageNet datasets show consistent gains while preserving the efficiency of RANet. Our code is available: https://github.com/jefersonseverino/ADC-RANet
Nov 7, 2026

Heterogeneous Asymmetric Contrastive Distillation

Knowledge distillation (KD) between architectural families, e.g., from a Transformer teacher to a CNN student, is difficult because the two networks learn different representations. Logits share a label space, but heterogeneous teachers differ in logit offset, scale, and sharpness, and their non-target classes are less semantically organized. Conventional logit distillation matches these values sample by sample, asking the student to imitate how another architecture expresses its predictions. We argue that a heterogeneous teacher should pass on the relational structure of its predictions, i.e., how it arranges samples relative to one another, and propose Heterogeneous Asymmetric Contrastive Distillation (HACD) to transfer this structure. HACD maps teacher and student logits to centered, temperature-softened probability directions on a shared unit hypersphere, which removes architecture-dependent offsets and lets the temperature control how sharply peaked outputs enter the representation while keeping informative non-target classes. Each student output must then identify its own teacher target within the batch. Because every teacher target also acts as a negative for the others, HACD computes the targets on a moderately augmented view and contrasts both a moderate and a strong student view against them. HACD adds no learnable parameters and needs one teacher pass per sample. It achieves the best or second-best accuracy on 11 of 12 heterogeneous pairs on CIFAR-100, without ground-truth labels, and on 14 of 15 pairs on ImageNet-1K, improving over the reported vanilla KD results by 8.82 and 1.33 percentage points on average. Checkpoints are available at https://github.com/BTAEYANG/HACD-Heterogeneous-Asymmetric-Contrastive-Distillation; the code will be released upon publication.
Nov 7, 2026

Understanding the Effect of Visual Attention in CNNs: A Decision-Space Alignment Perspective

Visual Attention enhances convolutional neural networks (CNNs) by modulating intermediate features, yet their effects on downstream decisions remain less understood. We present a decision-space framework that models multiplicative attention as an intermediate feature perturbation and derives a first-order prediction of its effect on the target-versus-competitor classification margin. We further introduce the Decision Alignment Score (DAS) to quantify the directional alignment between the perturbation and the local margin gradient. We evaluate the framework on CIFAR-100 using SimAM and CBAM across five popular CNN backbones, ResNet-18, ResNet-20, ResNet-50, ResNet-56, and MobileNetV2. The analysis examines prediction agreement, DAS margin correlations, and downstream activation-state changes. Across most final-attention configurations, predicted and actual margin changes show good agreement, while DAS exhibits positive correlations with actual margin changes. Single-attention experiments further show that prediction agreement and DAS margin correlations depend on attention placement, with stronger associations at later stages.
Nov 7, 2026

Evaluating classification and explainability in multimodal petrographic image analysis

Automated thin-section classifiers increasingly combine plane-polarised and cross-polarised light (PPL and XPL), two complementary views central to petrographic identification. Classification accuracy alone does not reveal whether model errors reflect mineralogical relationships, how predictions depend on each optical view, or whether attribution maps identify regions that influence model decisions. Here, we examine these complementary aspects of the released LITHOS dual-stream vision transformer using outputs from an earlier evaluation of 105,802 expert-annotated grains across 25 categories. Classification errors frequently followed the label taxonomy: 72.6% of misclassified plagioclase, microcline and sanidine grains were assigned the generic Feldspar label, while 87.1% of quartz-family grains received a label within that family, compared with 66.6% assigned their exact label. Replacing the XPL input with its per-channel spatial means produced a greater reduction in true-class probability than equivalent PPL replacement in 20 of 25 categories, but not in the three quartz categories. Five explainable artificial intelligence (XAI) attribution procedures generated visually distinct maps for identical inputs. In a superpixel-deletion analysis involving 30 grains from 29 thin sections, method-minus-random contrasts ranged from −0.043-0.043 to 0.0570.057, with all 41 simultaneous 95% confidence intervals including zero. Together, these findings show that error structure, modality sensitivity and attribution behaviour capture distinct aspects of model performance that cannot be established from classification accuracy or visual inspection of attribution maps alone.
Nov 4, 2026

Object-independent evaluation of small-data astronomical image classifiers: architecture, representation, and acquisition shortcuts

Reported performance in small astronomical image datasets can reflect object reuse, field-of-view differences, acquisition settings, and catalogue-derived properties as well as astronomical morphology. We designed \dataset{} to measure these effects in a four-class task comprising galaxies, globular clusters, nebulae, and open clusters. The dataset contains 370 catalogue-supported physical objects, each represented by three Digitised Sky Survey FITS bands, and uses a group-constrained split of 222 training, 75 validation, and 73 test objects. We compared a custom CNN--Transformer trained from scratch with ImageNet-pretrained ResNet-50 and ViT-B/16 under an adaptive full-field representation, a common angular field, and two central-region-masked controls, using three fixed seeds. In the one-pass object-independent test, mean balanced accuracy on the adaptive view was 0.773±0.0140.773\pm0.014, 0.874±0.0540.874\pm0.054, and 0.866±0.0230.866\pm0.023 for the custom model, ResNet-50, and ViT-B/16, respectively. No primary architecture contrast rejected its null after Holm correction; ResNet-50 and ViT-B/16 differed by 0.007 balanced-accuracy points (95% cluster-bootstrap interval [−0.048,0.062][-0.048,0.062]). Controls based on acquisition and catalogue metadata reached 0.721--0.735, and central-region-masked adaptive views reached 0.731--0.800. These results demonstrate generalisation to unseen physical objects, but also show that image context, acquisition settings, and catalogue-derived object properties carry substantial class information. In this regime, high accuracy cannot be interpreted as evidence that a classifier learned target morphology, and the comparison does not isolate architecture from pretraining.
Nov 3, 2026

Toward an Optimal Configuration for Tree Species Classification from Bark Images: A Leakage-Safe Ablation Study

Background: The automatic identification of tree species from bark photographs is of practical interest for forestry inventory and the wood-processing industry, yet the design choices that govern the accuracy of deep-learning approaches have not been studied systematically, and some prior results may be inflated by leakage of images of the same trunk across training and test partitions. Methods: A systematic ablation study is conducted on the public BarkNet 1.0 dataset (20 species) under the established tree-level, leakage-safe partitioning scheme. Using ConvNeXt-V2 backbones, the following factors are varied in turn and evaluated by five-fold cross-validation: patch size, the strategy by which per-patch predictions are aggregated into an image-level decision, the training regime (single- versus two-stage), backbone capacity, the whole-image baseline, and the number of patches available at inference. Results: Per-patch accuracy rises monotonically with patch size (from 75.6% at 160 px to 89.6% at 512 px), and image-level accuracy rises in step, plateauing near 94.9% around 512 px. Learned aggregation exceeds majority voting by at most 0.7 pp, an advantage that shrinks as patches grow and attention becomes near-uniform. Backbone capacity (pico to tiny) alters accuracy by less than a point, and two-stage training offers no advantage over one-stage. Whole-image classification (91.9%) trails patch-based aggregation, since it admits no bag to average over; accuracy increases with the number of patches at inference, saturating around 64. Conclusions: The study yields a simple recommended configuration: patches as large as possible while leaving enough to average over (around 512 px here), with a nano backbone. It also quantifies the second-order role of learned aggregation and model capacity, providing a reproducible reference point for practitioners and future research
Nov 3, 2026

LIPS-PC: Learning Illumination Patterns for Single-Pixel Classification

In single-pixel (ghost) imaging, known illumination patterns probe an object and a detector with no spatial resolution returns one number per pattern. Each measurement costs acquisition time and photons and, for ionizing probes, contributes to delivered dose. Motivated by automated radiographic defect screening, where a class label suffices, we introduce Learning Illumination Patterns for Single-Pixel Classification (LIPS-PC). The patterns are learned end-to-end while the measurement operator remains one fixed linear matrix, identical for every input. We parameterize it as a residual correction to a two-dimensional discrete cosine transform (2D-DCT), D=Dbase+ΔD=D_{\mathrm{base}}+Δ, and jointly train the correction and a compact classifier using cross-entropy, without a reconstruction objective. This yields a task-driven linear embedding through optimization. We evaluate LIPS-PC numerically on MNIST and Fashion-MNIST under a controlled protocol, then on RIAWELC, a real industrial X-ray weld-radiograph dataset. Across five budgets, M∈8,16,32,64,128M\in{8,16,32,64,128}, corresponding to sampling ratios from below 1% to 12.5%, the learned operator outperforms the best tested fixed transform by 9-15 percentage points at M=8M=8 on all three datasets. Its eight-measurement accuracy requires roughly 16-32 measurements with the best fixed transform, indicating a 2×2\times to 4×4\times saving at comparable accuracy. The advantage generally shrinks as MM increases, with a small reversal on RIAWELC at higher budgets. These results show that learning what to measure is especially valuable when measurements are scarce.
Nov 3, 2026

Statistical Machine Learning for Medicinal Plant Leaves Classification

Medicinal plants are usually identified using practitioners' experience, but reproducible, image based methods can complement this knowledge. We present MEDIPI (MEDIcinal Plant Identification), a statistical machine learning approach that classifies medicinal plant leaves from photographs. From each leaf image we extract 52 interpretable features covering shape, texture, color, and a new scagnostics category. We then classify leaves using a three level hierarchy: shape, then edge type, then species. We also release MedLEA, an open R package with 1,099 leaf images of 31 Sri Lankan medicinal plant species. On this dataset, a hierarchical Random Forest achieved the best overall accuracy (100%) in our comparison of hierarchical and non-hierarchical approaches and eight algorithms: Random Forest, XGBoost, Gradient Boosting, linear discriminant analysis, k-nearest neighbor, support vector machine, probabilistic neural network, and artificial neural network. Adding texture, color, and scagnostic features to shape features raised shape level accuracy from 90% to 97%. Linear discriminant analysis plots show how the trained model separates leaf shapes. Accuracy dropped sharply (23-50%) when training and test data came from different datasets, so generalization across datasets remains a challenge.
Nov 2, 2026

PriCBMs: Multi-granular Concept Bottleneck Vision-Language Models for Visual Privacy Recognition

Widespread image sharing on social media has increased the need for visual privacy recognition to identify inadvertently exposed privacy information. However, multi-label visual privacy recognition presents two challenges. First, privacy categories represent high-level semantic attributes and do not have a one-to-one correspondence with visual objects. Different visual objects may involve the same privacy category, while the same type of object may contain different types of privacy-sensitive information. In this setting, direct prediction methods generally lack explicit guidance for identifying category-relevant privacy cues, leaving this process implicit in the learned image-to-label mapping and potentially limiting classification performance. Second, privacy cues are often distributed across image regions and may not be adequately represented by whole-image features. To address these challenges, we propose Privacy Concept Bottleneck Models (PriCBMs). By introducing privacy-related concepts as intermediate semantic representations, PriCBMs first identify category-relevant privacy cues through visual-concept matching and then predict privacy categories based on concept responses. This process makes privacy predictions more interpretable. To improve the quality of these concepts, we design a concept selection strategy based on target-category relevance, cross-category discriminability, and semantic diversity. Meanwhile, we introduce a global-region modeling module to capture localized privacy evidence while retaining global context. Experiments on VISPR and FineViP show that PriCBMs outperform evaluated competitors across all four multi-label metrics. Ablation studies and concept-quality analyses support the effectiveness of the proposed design. Code is available at https://github.com/sun2293674491-dotcom/PriCBMs.
Nov 1, 2026

Systematic Evaluation of Robustness in Fine-tuned Models

Our study systematically evaluates the robustness of fine-tuned pre-trained models across different datasets and attack types, aiming to provide insights for researchers in selecting pre-trained models for image classification tasks while prioritizing robustness against potential attacks. Our work explores the influence of fine-tuning pre-trained models with different datasets and attack types on their robustness and overview of the appropriate pre-trained model for image classification based on factors such as computation resources, memory usage, and accuracy on clean data. We fine-tune the models with weights. Evaluate the models on data before any attack and after different attack cases. The study utilizes publicly available image datasets such as MNIST, CIFAR10, CIFAR100, and SVHN to assess the robustness of pre-trained models. The results indicate that the robustness of fine-tuned models varies across different data distributions and sample sizes, emphasizing the need for tailored selection based on specific task requirements. The findings also highlight the varying performance of ResNet50, DenseNet, and MobileNet in defending against attacks on different datasets. Overall, this research provides valuable insights into the robustness of fine-tuned pre-trained models and their suitability for real-world applications in image classification.
Oct 31, 2026

[Re] Pure Noise to the Rescue of Insufficient Data

We examine the main claims of Zada, Benou, and Irani (2021), which states that in an image classification task with imbalanced training data, (i) using pure noise to augment minority-class images encourages generalization by improving minority-class accuracy. This method is paired with (ii) a new batch normalization layer that normalizes noise images using affine parameters learned from natural images, which improves the model's performance. Moreover, (iii) this improvement is robust to varying levels of data augmentation. Finally, the authors propose that (iv) adding pure noise images can improve classification even on balanced training data. We implemented the training pipeline from the description of the paper using PyTorch and integrated authors' code snippets for sampling pure noise images and batch normalizing noise and natural images separately. We reproduced the main claims that (i) oversampling with pure noise improves generalization by improving the minority-class accuracy, (ii) the proposed batch normalization (BN) method outperforms baselines, (iii) and this improvement is robust across data augmentations. Our results also support that (iv) adding pure noise images can improve classification on balanced training data. However, additional experiments suggest that the performance improvement from OPeN may be more orthogonal to the improvement caused by a bigger network or more complex data augmentation.
Oct 31, 2026

Adversarial robustness does not entail human-aligned categorization

Adversarial training can improve some forms of correspondence between artificial and biological vision, but whether robust classifiers acquire human-aligned classification behavior remains unclear. We generated 1,008 controversial stimuli by optimizing latent inputs to a pretrained GAN to elicit high-confidence disagreements between pairs of robust ImageNet classifiers. We compared four adversarially robust ImageNet classifiers - ResNet-101, ViT-B, ConvNeXt-L, and Swin-L - with two-alternative category judgments from 60 participants. We first evaluated each classifier on stimuli optimized to elicit disagreement with another model. Across classifiers, accuracy in predicting individual participants' single-trial category choices was near chance (48.6-51.3%), whereas majority judgments from other participants predicted responses with 85.0-86.8% accuracy. An analysis incorporating model prediction strength and participants' reported confidence found weak positive model-human correspondence for the two convolutional models. On stimuli optimized for disagreement between other model pairs, all four robust classifiers predicted human responses above chance but less accurately than predictions based on other participants' majority judgments; the same pattern held for 17 additional RobustBench models not used in stimulus generation. Across these 17 models, mean model-participant agreement was 67.3%, versus 81.3% mean participant-participant agreement on the same images. These results identify model-specific failures of human alignment and a model-human mismatch that persists across robust classifiers and substantially exceeds disagreement among human observers. Our findings show that adversarial robustness alone does not guarantee human-aligned categorization and suggest that robustness and human alignment should be evaluated as distinct properties of visual recognition systems.
Oct 30, 2026

OT-CAL: Optimal-Transport-Guided Adaptive Sample Selection for Efficient Visual Learning

Deep neural networks for image recognition are typically trained using mini-batches, with a forward pass to compute predictions and losses followed by a more computationally expensive backward pass to update model parameters. However, not all training samples contribute equally to learning, suggesting that computation can be reduced by performing backward passes only on informative samples. The challenge is that sample informativeness changes throughout training: difficult samples may be valuable early on, improving samples become important later, and class diversity becomes increasingly relevant toward the end. Existing methods often rely on fixed selection rules or computationally expensive per-sample gradient calculations. We propose OT-CAL, an adaptive data selection method that automatically adjusts its sampling preferences during training. OT-CAL represents each sample using five informativeness features, including difficulty, learnability, and uncertainty. A curriculum vector describes the currently preferred sample characteristics, and optimal-transport-based matching assigns selection weights according to how closely samples match this target. The curriculum vector is then updated using the selected samples, forming a feedback loop that adapts without a manually defined schedule. Experiments on three standard visual learning benchmarks show that, when selecting 50% of each mini-batch, OT-CAL improves performance by 1.3--2.1 points over AdaSelection while achieving 1.18--1.24 times faster training than full-data training.
Oct 30, 2026

BDNet: Fast-slow thinking is all you need

We present BDNet, a hierarchical image-classification backbone built around fast image features and slow statistical references. Its baseline deformation operator (BDO) transforms the deviation of a feature from a stage-specific reference through a low-rank projection and a learned channel scale. Rather than computing pairwise token similarity, the branch operates directly on baseline-relative residuals with complexity linear in the token count. Each of four reference buffers accumulates detached stage statistics during training and remains frozen at inference. A four-stage pyramid with depths (2,2,6,2)(2,2,6,2) and widths (96,192,384,768)(96,192,384,768) yields a 20.70M-parameter classifier at 224×224224\times224 resolution. We organize an ImageNet-1K study around clean classification, eight controlled ablations, and Gaussian-noise, contrast, and brightness perturbations. The central hypothesis is that a slow reference can improve the robustness of a compact representation while retaining a simple tensor implementation. We distinguish this training-time statistical mechanism from input-dependent attention and make its linear inference behavior explicit.
Oct 29, 2026

Uncertainty-aware Feedback Loop Diagnosis: integrating reinforcement learning and conformal prediction for reliable wafer defect classification

Semiconductor manufacturing is a complex, multi-step process that requires both high precision and low operating costs. Therefore, detecting defects in a high-volume setting is an important component of wafer map analysis for reliable production. However, most state-of-the-art deep learning approaches require a full scan of the surface, which hinders the development of high-throughput inspection systems. To address this problem, we propose Uncertainty-aware Feedback Loop Diagnosis (UFLD), a novel algorithm that employs reinforcement learning for adaptive measurement acquisition and provides a prediction uncertainty level after each step, stopping the process once sufficient information is obtained. Unlike existing methods that rely on ensemble models or auxiliary confidence estimation deep learning models, UFLD utilizes a more parameter-efficient conformal prediction framework to determine the optimal stopping time. The proposed UFLD reduces acquisition time by about 40%40\% and achieves 98.4%98.4\% inspection accuracy, without requiring an additional learned confidence-estimation network. We validated our approach on the MixedWM38 dataset, demonstrating its advantages over existing adaptive acquisition procedures and standard convolutional neural networks.
Oct 28, 2026

Automatic Iconographic Annotation of Medieval Illuminations: Comparing Generative Vision-Language Models and Supervised Discriminative Classifiers

The large-scale digitisation of medieval manuscripts creates a growing need for automated iconographic indexing. We evaluate two complementary paradigms for this task on 9,735 full-folio images annotated with 2,333 TIMEL concepts: generative Vision-Language Models (VLM) and supervised discriminative classifiers (SDC). The VLM experiments evaluate fine-tuning, structured JSON decoding, taxonomybased candidates generation, and deterministic reranking. The discriminative experiments evaluate flat, hierarchy-aware, and patch-attention heads over a frozen DINOv3 encoder. Our results highlight key functional differences: ML-Decoder (SDC) achieves the strongest overall retrieval performance (precision@1 of 0.696, MRR of 0.806), with loss calibration (Asymmetric Loss) serving as the primary performance lever. While SDCs provide superior precision and reliability within a closed vocabulary, VLMs offer strong semantic generalization and broad taxonomic coverage, particularly through structured JSON targets and verify-calibrate reranking (hierarchical F1 of 0.559). Furthermore, model behavior strongly depends on concept abstraction: SDCs excel at perceptually grounded descriptive concepts, whereas VLMs provide useful hypotheses for complex interpretative scenes. Finally, we find that enforcing explicit hierarchy within model architectures limits coverage compared to spatial feature aggregation. Overall, rather than competing approaches, SDCs and VLMs emerge as complementary components for future hybrid indexing systems combining structured retrieval with exploratory annotation assistance. As a methodological contribution, we define the soft F1 metric to provide a taxonomy-aware evaluation that credits semantically close predictions based on lowest common ancestors.
Oct 28, 2026

TreemapMix: Dirichlet-Controlled Multi-Image Augmentation for Probability and Ordinal Supervision

Data augmentation improves visual recognition by exposing models to synthetic training examples that encourage generalization beyond the observed data. Mixed-sample methods extend this idea by combining multiple images. However, many common approaches either mix only pairs of images or rely on fixed small layouts, limiting control over source count, visible area proportions, and induced supervision. We propose TreemapMix, a distribution-controlled multi-image augmentation that samples source weights from a Dirichlet distribution and assigns them to area-proportional regions using a treemap partition. The resulting known region areas support two complementary objectives. TreemapMix with soft cross-entropy (SCE) treats areas as soft class-probability targets, while TreemapMix with a Plackett-Luce (PL) objective converts areas into ordinal supervision. On ImageNet-1K under matched training recipes, TreemapMix-SCE achieves substantially lower calibration error than evaluated mixed-sample baselines, while TreemapMix-PL achieves the strongest Top-1 accuracy. Transfer experiments on object detection and instance segmentation suggest that TreemapMix-pretrained features remain useful beyond classification. These results show that controlled multi-image composition can expose the same region-area information as either probability-matching or rank-matching supervision. Code is available at https://github.com/Ejafa/treemapmix.
Oct 27, 2026

You Don't Need Aligned Representations: Knowledge Distillation via Random Prototype Spaces

A widely held assumption in knowledge distillation is that teacher and student representations must occupy a compatible or explicitly aligned feature space. We propose RPKD (Random Prototype Knowledge Distillation), which discards this assumption by comparing both networks against a shared bank of randomly initialized, frozen prototype vectors. Alignment is therefore carried out in an external reference space rather than in the teacher's native feature coordinates: the teacher side is projected by fixed random maps that carry no trainable parameters, and only the student adapts. Two branches handle the transfer: a logit branch aligning global prototype similarity distributions and a feature branch matching spatially-resolved prototype activation maps. Both operate within the same fixed prototype space, which can be interpreted as a random feature embedding that approximately preserves inter-sample similarity structure, so the prototype bank and the distillation objective stay unchanged across teacher-student pairs while only the input projections are instantiated per architecture. In low-category settings, decoupling the prototype vocabulary from the task label space yields richer supervision than the label space alone, an advantage absent from logit-based methods, whose supervisory signal collapses as class count shrinks. This advantage is especially relevant in real-world scenarios where the label space is typically constrained. Extensive experiments demonstrate the effectiveness of RPKD, achieving a maximum gain of +7.98% over OFA on CIFAR-100, +12.44% on ImageNet-100, and +1.28% on ImageNet-1K. Source code is available at https://github.com/HamzaEtcibasi/RPKD
Oct 27, 2026

Subspace Clifford Networks

In this work, we propose Subspace Clifford Networks, a reformulation that replaces global cyclic rolling with grouped subspace interactions. Instead of sampling sparse cyclic diagonals across the entire DD-dimensional channel space, we project the features into an aligned space RD~=⨁g=1GVg\R^{\tilde{D}} = \bigoplus_{g=1}^{G} V_g partitioned into GG low-dimensional groups Vg≅RkV_g \cong \R^k with k∈{2,3,4}k \in \{2, 3, 4\}. Within each kk-dimensional group, we compute all (k2)\binom{k}{2} wedge coordinates and concatenate them with SiLU⁡\operatorname{SiLU}-activated coordinate-wise products, while sandwiching the grouped interaction between learnable 1×11\times 1 linear projections to enable cross-group channel mixing. This design reduces intermediate feature expansion from 2∣S∣D2|\mathcal{S}|D (typically 10D10D for 5 shifts) down to 1.5D~1.5\tilde{D} (k=2k=2), 2.0D~2.0\tilde{D} (k=3k=3), or 2.5D~2.5\tilde{D} (k=4k=4). Furthermore, exploiting the small, fixed group size k∈{2,3,4}k \in \{2, 3, 4\}, we implement a fused CUDA kernel that evaluates the forward interaction and a recomputation-based analytical backward pass using thread-local variables without allocating global-memory buffers for elementwise intermediates. Finally, we extend the backbone from an isotropic columnar layout to a multi-resolution Hierarchical Clifford Network with configurable per-stage subspace dimensions. Experiments on CIFAR-100 show that the proposed models substantially reduce parameter count and training time across both regimes, while the hierarchical variants also improve classification accuracy over the evaluated shift-based hierarchical baselines.
Oct 27, 2026

Ground-to-Satellite Knowledge Distillation for Crop Classification under Temporal and Geographic Shift

Continuous and up-to-date crop type mapping is essential for agricultural monitoring but is often hampered by limited reference data. The growing availability of ground-level field photographs offers a promising way to address this challenge. When combined with satellite time series, these images provide complementary perspectives on the same parcel, capturing crop characteristics from distinctly different perspectives. However, field photographs are not always available for each satellite pixel at inference time, limiting their potential in scalable crop type mapping. We investigate whether ground photographs can provide privileged supervision for crop classification built on pretrained satellite embeddings. A photo teacher based on frozen CLIP features supervises classifiers trained on AlphaEarth and TESSERA representations; photographs are used only during training. On a paired field photo and satellite imagery dataset with 17 crop classes, knowledge distillation consistently improves over satellite-only classification; for frozen embeddings the margin grows under temporal shift, reaching up to +5.5 macro-F1 points, and widens in geographic holdouts (+3.7 vs +3.0 in-distribution). These results show that ground-view supervision can complement satellite embeddings and improve their robustness without changing inference requirements.
Oct 27, 2026

Tail Cues, Principal Corrections: Plug-and-Play Rectification for Open-Set Test-Time Adaptation

Open-Set Test-Time Adaptation (OSTTA) operates on non-stationary target streams where covariate and semantic shifts coexist. Existing TTA methods often obtain adaptation signals by updating model parameters or maintaining mutable memories, which can be costly and vulnerable to unknown samples. We present TALOR, a lightweight, backward-free rectification module that uses the source linear head as a geometric anchor. The head-induced basis decomposes each normalized feature into high-gain principal coordinates and low-gain tail coordinates. TALOR estimates a soft-routed affine correction in the principal subspace, combining a regime-level principal bias with a tail-conditioned slope that captures sample-level tail-to-principal coupling. Using weighted ridge regression, it subtracts this drift estimate from the principal coordinates, reconstructs the feature and feeds it into the source head. On the ImageNet-C benchmark with six csOOD datasets, TALOR outperforms the next-best UniEnt by 2.2% H-score while running 2.5x faster and using only 12% of its GPU memory; as a post-correction plug-in, it further boosts COME to 68.4% H-score with only 2% additional memory overhead.
Oct 26, 2026

ACSAN: All-Convolutional Self-Attention Network Integrating Convolutional Operations with Vision Transformer

In the past few years, Vision Transformer (ViT) has achieved competitive performance in the field of computer vision. However, there still remain some problems. For instance, ViTs rely on large-scale training data since they simply contain a large number of trainable parameters, which slows down their learning efficiency and limits their applicability in resource-constrained scenarios. In this paper, we propose a novel deep learning model, namely All-Convolutional Self-Attention Network (ACSAN), which applies convolution operations to replace fully-connected matrices multiplications in the ViT framework. To implement ACSAN, we firstly design Dynamic Convolutional Self-Attention (DCSA) as the basic module, which employs multi-scale dilated convolutions to generate Query, Key, and Value. We define spatial self-attention and channel self-attention to generate attention features respectively. The two attention features are finally merged as the output of the DCSA module. Based on DCSA, we construct the basic ACSAN block. Specifically, we use DCSA to replace the regular self-attention operations, and use convolution-based Equivalent Convolutional Fully Connected (EqConvFC) modules to replace MLPs. To evaluate the performance, we train ACSAN on CIFAR-10, CIFAR-100 and ImageNet-1k datasets from scratch. Experimental results show that ACSAN achieves good classification accuracy (the best variants achieve 93.52% on CIFAR-10, 73.38% on CIFAR-100, and 83.3% on ImageNet-1k). Moreover, in comparison with standard ViT and other state-of-the-art methods, ACSAN uses fewer parameters to achieve comparable or better accuracy.
Oct 26, 2026

Zero-Shot Multi-Class Wafer Map Classification with Rule-Ordered Prompts for Large Multimodal Models

Wafer-map defect classification often requires labeled data and retraining when production conditions change. We present a training-free six-class framework that combines connected-component highlighting with morphology-, location-, and exclusion-aware prompts for a frozen large multimodal model. On Mixed-WM38 (6,866 maps), multi-region prompting achieves 92.9% macro and 93.8% overall accuracy. On WM811K (9,000 maps), it achieves 79.1% and 81.3%, respectively. Eight alternative rule orders achieve 90.1-91.9% overall accuracy, compared with 93.8% for the selected order.
Oct 24, 2026

Component Pooling Trained Across Orientations for Few-Shot Remote Sensing Scene Classification

Few-shot remote sensing scene classification (FSRSSC) aims to recognize novel scene categories from only a few labeled images. Most FSRSSC methods summarize the feature map of an image by global average pooling, which treats the few objects that define an aerial scene and the large background around them alike, and they rarely consider that aerial images have no canonical orientation. To address these issues, we propose RoCoP, a parameter-free component pooling network trained across orientations. RoCoP consists of a feature extractor, a non-negative decomposition module (NDM) and a component scoring module (CSM). NDM factorizes the local descriptors of each image into a few additive components with a deterministic initialization and a single multiplicative update, so that the positions of one object are grouped together. CSM weights each component by its activation magnitude and converts the weights into a soft mask for pooling. Because both modules are invariant to permutations of feature positions, the only rotation-sensitive part of the pipeline is the backbone, which we train to classify randomly rotated views of every image. Experiments on NWPU-RESISC45, AID and UC Merced show that component pooling and rotated-view training are complementary: together they improve a strong transfer-learning baseline by about 2.5 points in 5-way 1-shot accuracy on NWPU-RESISC45 and AID, without adding parameters at inference. Under an identical protocol, RoCoP outperforms re-implemented metric-based and task-adaptive methods, and NDM and CSM also improve existing methods when plugged into them. We further identify the regime in which the method does not apply, namely very small base sets.
Oct 23, 2026

Automated identification of ivory carvings using machine learning: A proof-of-concept approach to detecting the illegal wildlife trade

Despite a ban on the international trading of recently poached ivory, elephants are still killed for their tusks. Ivory and its products are mainly sold in online marketplaces. However, due to the huge volume and variety of adverts and the similarity of ivory with other materials that can be legally sold, it is hard for enforcement officials to identify illicit trading and prosecute sellers. Existing techniques for identifying elephant ivory, such as Fourier-transform infrared spectroscopy and microscopy, can reliably distinguish ivory from other materials, but require physical access to the item and cannot be applied to photographs alone. As a result, these methods are unsuitable for screening the large volume of images posted in online marketplaces, leaving a gap for automated, image-based identification tools. In this proof of concept work, we investigate the potential of machine learning techniques for image analysis to automatically distinguish ivory objects from bone. We employ a semantic segmentation framework using a U-Net architecture with a ResNet34 backbone, leveraging transfer learning on an augmented dataset of 541 original images. Using a dataset of images of ivory and items resembling carved ivory, several models were trialled. On a distinct test set, the classifier achieved an Area Under the Curve (AUC) of 0.81 and an overall classification accuracy of 81% on confident samples. The paper presents the best-performing model as a proof of concept for developing a system that can effectively identify ivory and its products when indicative features of elephant ivory are present in an image. This work provides initial steps towards a machine learning-based framework designed to advise law-enforcement in detecting illegal elephant ivory trade.