Growing volumes of plastic waste highlight the need for reliable sorting technologies. This study examines whether lightweight YOLO object detectors can distinguish polyethylene terephthalate (PET) bottles from high-density polyethylene (HDPE) bottles, two common packaging materials that must be separated for recycling. Two public PET and HDPE datasets were augmented offline to simulate the visual variation typical of industrial sorting lines, and eleven compact YOLO models–nano, tiny and small versions across versions 7–12–were fine-tuned on these data. Their accuracy was quantified using mean average precision (mAP) at multiple intersection-over-union thresholds and precision, recall and F1 scores; their efficiency was evaluated by inference time, size and floating-point operations. All models achieved mAP@0.5 above 99.2% and F1 scores exceeding 98% on test data. Among them, YOLOv11n delivered the best trade-off between accuracy and speed, it achieved mAP@0.5:0.95 of 93.7%, processed 640 \(\times\) 640 pixel images in around 38 ms on a CPU, and required just 5.5 MB of memory. These findings suggest that accurate polymer discrimination is possible on modest hardware, enabling resource-constrained recycling facilities to reduce mis-sorting, trim costs and support a circular economy.
Contamination in recycling streams represents one of the most pervasive challenges confronting material recovery facilities (MRFs) globally. Among the various contamination sources in recycling streams, liquid contamination from PET bottles presents particularly severe challenges due to its capacity to spread throughout commingled materials. Object detection using neural networks enables detection at the collection stage of single or mixed recycling streams, allowing for targeted application in the early stage of the recycling cycle. YOLO (you only look once) models and other object detection models are beneficial due to their speed and accuracy in detecting multiple objects at once. This study aimed to design a model to detect contaminated PET bottles in real time. Several YOLO variations and model sizes were trained on a custom dataset with 7130 images. YOLOv8l achieved the highest performance, with mAP@0.5:0.95, mAP@0.5, precision, recall, and F1 score values of 89.7%, 93%, 89%, 88%, and 88%, respectively.
Ensuring the integrity of eggshells across poultry supply chains is vital to minimize economic losses and mitigate the risk of pathogen ingress. Manual visual inspection remains inconsistent, labor-intensive, and increasingly inadequate for modern production volumes. This study conducts a comprehensive comparative evaluation of lightweight deep-learning models for automated eggshell defect detection, assessing their suitability for deployment on resource-constrained augmented-reality (AR) smart glasses. A mixed dataset was assembled by combining 186 privately collected images of brown cage-free eggs with 794 publicly available images of primarily white eggs. Ten architectures, comprising nine lightweight detectors across the YOLOv5, YOLOv8, YOLOv9, and YOLOv10 families alongside a transformer RT-DETR-l baseline, were benchmarked. The YOLOv8n model achieved the highest detection accuracy with a 5-fold cross-validated mAP@0.5:0.95 of 91.2%, demonstrating no statistically significant difference relative to RT-DETR-l. However, YOLOv10n provided the optimal accuracy-efficiency balance for AR deployment as it delivered a mAP@0.5:0.95 of 89.5% and a proxy CPU inference time of 34.2 ms, while an INT8-quantized TFLite export reduced its memory footprint to 2.59 MB. By providing a mixed-egg dataset and a thorough benchmark of lightweight detectors, this comparative study identifies YOLOv10n as the most suitable architecture for real-time AR-based inspection. These findings advance precision agriculture by outlining an efficient, empirically validated pathway toward hands-free, AI-driven quality control.
Manual sorting of date fruits is time-consuming and labor-intensive, while automated optical sorting lines remain costly for small- and medium-sized enterprises (SMEs). This study presents a deployment-oriented benchmark of five lightweight object detectors for single-fruit date-variety recognition, namely the Convolutional Neural Network (CNN)-based YOLOv8, YOLOv10, YOLOv11, and YOLOv12 Nano variants and the Transformer-based D-Fine-Nano, with RT-DETR-Large included as a high-capacity accuracy reference. Models were evaluated on 1,658 controlled-environment images of nine commercial varieties using grouped 5-fold cross-validation, which kept identified repeat photographs of the same fruit within the same cross-validation fold. Latency was measured on an NVIDIA T4-class edge GPU/industrial-PC accelerator, and the CNN detectors were also exported to Open Neural Network Exchange (ONNX) and TensorRT half-precision (FP16) engines. All Nano models met the 30 frames-per-second threshold for model-side detection latency under the tested conditions and achieved high accuracy, while a Friedman–Nemenyi analysis detected no statistically significant accuracy difference among the lightweight models. Under TensorRT FP16, YOLOv10-Nano achieved the lowest model-side end-to-end latency of 6.27 ms with no observed accuracy degradation. YOLOv8-Nano attained the lowest inference-stage latency of 3.31 ms, but its mean average precision (mAP@50–95) decreased by 1.2 percentage points after TensorRT FP16 conversion. YOLOv12-Nano had the numerically highest mean cross-validation mAP@50–95 and the lowest fold-to-fold variability, albeit at higher latency. These findings characterize model-level detector behavior rather than a complete sorting line. Within that scope, YOLOv10-Nano represents one practical configuration for a T4/TensorRT FP16 pipeline, while the other models offer different speed, stability, and deployment trade-offs, providing SMEs with a well-characterized starting point for the vision component of affordable date-sorting systems.
Background/Objectives: Smartphone-acquired throat images could support accessible pharyngitis assessment, but the performance and computational efficiency of lightweight models on such images require further evaluation. This study compared five lightweight models for predicting physician-consensus bacterial versus nonbacterial labels from smartphone throat images. Methods: We analyzed PGUPharyngitis, a public dataset of smartphone throat images collected in general practice. The final cohort comprised 635 patients, of whom 105 (16.5%) had bacterial physician-consensus labels. The reference outcome was physician consensus based on images and clinical information rather than microbiological testing. Five ImageNet-pretrained lightweight classifiers—four convolutional neural networks (CNNs; MobileNetV3-Small, EfficientNet-B0, ShuffleNetV2, and YOLO26n-cls) and one CNN–Transformer hybrid (MobileViT-XS)—were fine-tuned using patient-level five-fold cross-validation. Performance was assessed using area under the precision–recall curve (PR-AUC) and area under the receiver operating characteristic curve (AUROC), alongside computational efficiency. Device-only and clinical-variable logistic-regression models provided non-image comparisons. Results: Mean PR-AUC across folds ranged from 0.317 to 0.372 and AUROC from 0.663 to 0.708, with no clearly superior architecture. Bacterial-label prevalence differed by acquisition device (Samsung 9.2% versus Xiaomi 26.9%); the device-only and clinical-variable comparators achieved AUROCs of 0.650 and 0.598, respectively. At validation-selected high-sensitivity thresholds, the models flagged 65–78% of patients, with positive predictive values of only 19–22%. All models contained fewer than five million parameters, making them compact candidates for future evaluation on resource-constrained devices. Conclusions: Lightweight models showed moderate performance against physician-consensus labels. However, acquisition device was associated with the reference labels, image-model AUROCs were not clearly distinguishable from that of the device-only comparator, and device-related confounding cannot be excluded. Because performance was measured against these labels and positive predictive values were low, the findings do not establish diagnostic accuracy or clinical applicability, which require validation against microbiological reference standards in independent clinical cohorts.