Aug 2026· Advances in Research· Vol 27, pp. 67-90· 0 citations
TL;DR
Evaluation in digital image processing is argued for evaluation that is task-aware, distribution-aware and resource-aware, with uncertainty, fairness, calibration and reproducibility treated as core properties rather than optional add-ons.
Abstract
Digital image processing has evolved from deterministic operators designed around explicit assumptions to data-driven systems that learn representations, priors and decision rules from large image collections. This transition has improved performance in restoration, segmentation, compression and visual interpretation, but it has also complicated evaluation, reproducibility and deployment. This critical narrative review examines the field as a connected pipeline rather than as a catalogue of algorithms. It considers image formation and degradation, classical spatial and variational methods, non-local and sparse modelling, convolutional neural networks, vision transformers, diffusion models and emerging foundation models. It also analyses how these approaches are translated into medical imaging, Earth observation, industrial inspection, agriculture and multimedia systems. The evidence indicates that classical methods remain valuable when physical assumptions are credible, data are scarce or interpretability and predictable failure behaviour are important. Learned models are strongest when training and deployment distributions are aligned and task-specific data are sufficiently representative, while hybrid physical and learned approaches can reduce sample requirements and constrain implausible outputs. Transformers and diffusion models broaden context modelling and generative capability, although their computational cost, dependence on pretraining and uncertain behaviour under distribution shift limit universal claims of superiority. Across applications, benchmark gains frequently exceed improvements demonstrated in prospective or operational settings. Image quality metrics, dataset design and reporting practices often fail to capture clinically, environmentally or industrially consequential errors. The review therefore argues for evaluation that is task-aware, distribution-aware and resource-aware, with uncertainty, fairness, calibration and reproducibility treated as core properties rather than optional add-ons. Future progress is likely to depend less on isolated architectural novelty than on credible image-formation models, representative data, robust adaptation, efficient computation and validation against domain outcomes.
Tremendous numbers of images and videos are generated throughout the life cycle of construction projects, offering rich visual data that, if analysed effectively, can support automation in progress monitoring, safety management, quality control, and productivity tracking. Within this context, image classification plays a foundational role: it is the supervised task of mapping an image to one of K semantic categories by learning from labelled examples, requiring the extraction of discriminative visual structure and the production of calibrated class predictions. Despite major advances from handcrafted feature pipelines to Convolutional Neural Networks (CNNs), fair comparison across methods remains challenging due to inconsistent preprocessing, dataset splits, training budgets, and reporting standards. This paper introduces a unified and fully reproducible benchmarking framework that standardises preprocessing, enforces fixed train–test splits, aligns training configurations, and evaluates models using a consistent multi-metric suite. Under identical experimental conditions, we benchmark four paradigm families: classical machine learning (ML) models, handcrafted feature-based methods (Histogram of Oriented Gradients (HOG), Local Binary Patterns (LBP), Bag-of-Visual-Words (BoVW) [Formula: see text] Support Vector Machine (SVM)), and CNNs (LeNet-5, ResNet-18), across three datasets (MNIST, Fashion-MNIST, CIFAR-10). We formalise image classification mathematically as a supervised learning problem and assess performance using accuracy, macro-averaged [Formula: see text]1-score, and macro-averaged ROC–AUC, complemented by learning-curve and robustness analyses. Results show that ResNet-18 achieves the strongest performance on all datasets (98.95%, 90.27%, 72.90%), SVM with a Radial Basis Function (RBF) kernel remains competitive on grayscale benchmarks, and hand-crafted pipelines degrade sharply on CIFAR-10. ROC–AUC analysis further highlights calibration–accuracy discrepancies, such as Naïve Bayes achieving high AUC but modest accuracy on MNIST. These findings emphasise the importance of task-aware model selection and demonstrate that compute-aligned, multi-metric benchmarking is essential for fair and reproducible cross-paradigm evaluation. All CNN results are reported under a 50-epoch training budget, which represents a lower bound on converged performance; future extensions will incorporate domain-specific datasets and pre-trained architectures to broaden practical applicability.
Yasir Hamid, N. Ababneh, A. Abasi et al.· Journal of Information &...· 0 citations
In medical imaging, it is common to use learned perceptual image patch similarity (LPIPS) to compare images semantically in feature space. Although backbones pretrained on natural images are widely used for LPIPS computation, B-mode ultrasound images possess distinct speckle patterns and acoustic-specific image statistics that are fundamentally different from natural images and even from other images in radiology. Consequently, we propose that domain-specific models are needed to measure perceptual similarity in ultrasound data, a finding which is not necessarily the case for other imaging modalities. We compare LPIPS metrics across downstream tasks like classification, segmentation and reconstruction using natural image, medical generalist and ultrasound backbone models and show that selection of LPIPS backbone is a non-trivial design choice. In particular, the ultrasound backbone models were more correlated with downstream performance of supervised models than classical and natural image models, and optimization of the LPIPS loss with an ultrasound backbone achieved a strong balance between reconstruction quality and realism. Our code is available at https://github.com/talg2324/UltraPIPS and introduces the UltraPIPS library, a set of LPIPS metrics based on the open-source foundation models analyzed in this paper.
The optimization objective in image dehazing is fundamentally shifting from deterministic pixel mapping to high-dimensional probability distribution modeling. This survey organizes the field according to three optimization paradigms: physical prior-guided deterministic mapping, end-to-end feature reconstruction, and generative distribution alignment. Our synthesis yields three main findings. First, single-pass CNN, Transformer, and state-space models remain attractive for latency-sensitive applications, although pixel-wise objectives can suppress high-frequency and perceptually plausible details. Second, GAN- and diffusion-based methods often report improved perceptual or distributional quality when evaluated using LPIPS or FID, but iterative sampling and weak physical anchoring increase computational cost and the risk of structurally inconsistent details under dense haze. Third, heterogeneous datasets, image resolutions, evaluation protocols, and incomplete perceptual reporting do not currently support a controlled quantitative comparison of hallucination rates across architectures. We therefore analyze the role of atmospheric scattering constraints in anchoring generative trajectories and advocate physics-consistent, task-driven evaluation. Future work should develop unified hallucination benchmarks and efficient, physically constrained generative models for edge and safety-critical deployment.
An efficient aligned kernel network (AKNet) is proposed, which innovatively employs super-large convolution kernels to capture global receptive fields with minimal computational overhead, effectively mimicking the long-range dependency modeling of transformers.
Wan Li, Xiao-Lin Zhang· The Visual Computer· 0 citations