This research describes the use of high-performance computing (HPC) and deep learning to create prediction models that could be deployed on edge AI devices equipped with camera and installed in poultry farms. The main idea is to leverage an existing IoT farming platform and use HPC offline to run deep learning to train the models for object detection and object segmentation, where the objects are chickens in images taken on farm. The models can be ported from HPC to edge AI devices to create a new type of computer vision kit to enhance the existing digital poultry farm platform. Such new sensors enable implementing functions such as counting chickens, detection of dead chickens, and even assessing their weight or detecting uneven growth. These functions combined with the monitoring of environmental parameters, could enable early disease detection and improve the decision-making process. The experiment focused on Faster R-CNN architectures and AutoML was used to identify the most suitable architecture for chicken detection and segmentation for the given dataset. For the selected architectures, further hyperparameter optimization was carried out and we achieved the accuracy of AP = 85%, AP50 = 98%, and AP75 = 96% for object detection and AP = 90%, AP50 = 98%, and AP75 = 96% for instance segmentation. These models were installed on edge AI devices and evaluated in the online mode on actual poultry farms. Initial results are promising, but further development of the dataset and improvements in prediction models is needed.

This research explores the role of synthetic data in enhancing the accuracy of deep learning models for automated poultry farm management. A hybrid dataset was created by combining real images of chickens with 400 FLUX.1 [dev] generated synthetic images, aiming to reduce reliance on extensive manual data collection. The YOLOv9 model was trained on various dataset compositions to assess the impact of synthetic data on detection performance. Additionally, automated annotation techniques utilizing Grounding DINO and SAM2 streamlined dataset labeling, significantly reducing manual effort. Experimental results demonstrate that models trained on a balanced combination of real and synthetic images performed comparably to those trained on larger, augmented datasets, confirming the effectiveness of synthetic data in improving model generalization. The best-performing model trained on 300 real and 100 synthetic images achieved mAP = 0.829, while models trained on 100 real and 300 synthetic images reached mAP = 0.820, highlighting the potential of generative AI to bridge data scarcity gaps in precision poultry farming. This study demonstrates that synthetic data can enhance AI-driven poultry monitoring and reduce the importance of collecting real data.

Highland pastures in Montenegro, known as katuns, are seasonal settlements without grid power or network coverage and which are located where conventional monitoring is unfeasible. This study presents a solar-powered, off-grid system for livestock and environmental monitoring. It integrates, into a single portable unit, a solar power station, an edge-AI computer, a camera, environmental sensors, a LoRaWAN gateway, and a cellular router for backhaul. All parts are pre-wired in a modular enclosure, deployable by one operator in under 30 min. Data are fed to the agroNET farm-management platform and a purpose-built mobile web application; livestock detection runs on-device using a model from our earlier work. The system was evaluated at three sites, including a highland katun near Žabljak (~1450 m), under a two-phase energy-measurement protocol. During field logging it drew ~75 W on average against ~125 W solar input—a measured surplus that is used to recharge the battery—with a daily monitoring load of ~1560 Wh. The four-panel array’s nameplate potential in summer is an estimated ~3700 Wh/day, indicating substantial headroom relative to the measured load. At 80% depth of discharge the battery gives ~20 h autonomy, and the detection pipeline ran continuously, processing ~10,000 frames at under 3 s latency. The results demonstrate the feasibility of off-grid precision livestock farming, reaching TRL 6.

Introduction/Objective: Retinal fundus photography captures microvascular and neuroretinal features reflecting systemic health. While Deep Learning (DL) has enabled extraction of cardiometabolic and aging biomarkers from retinal images, its application to oncologic disease status discrimination remains largely unexplored. This study evaluated whether DL analysis of retinal fundus photographs can discriminate individuals with colorectal cancer (CRC) from those without CRC (non-CRC) in a curated UK Biobank case-control setting. Methods: This retrospective imaging study used 1,600 UK Biobank fundus photographs (800 CRC, 800 non-CRC). Images underwent grayscale conversion, intensity normalization, and resizing with minimal preprocessing. Data were split at the participant level into training, validation, and test sets (70/15/15). An ImageNet-pretrained EfficientNet-B3 convolutional neural network was fine-tuned end-to-end for binary classification. Model performance was assessed using Area Under the Receiver Operating Characteristic Curve (AUROC). Prevalence-adjusted predictive values were estimated to approximate real-world classification performance. Results: On the independent test set, the model achieved AUROC = 0.945, accuracy = 87.08%, macro-F1 = 87.07%, sensitivity = 90.00%, specificity = 84.17%, and precision (PPV in the balanced test set) = 85.04%. When recalibrated to population prevalence, PPV decreased to 2.8% at 0.5% prevalence (and 6.5% at 1.2% prevalence), whereas NPV remained high (99.94% and 99.86%, respectively). Discussion: High discrimination achieved from retinal images alone supports the presence of CRC-associated retinal signatures. A covariates-only baseline using age and vascular/heart disease showed only modest discrimination (AUROC ≈ 0.63), making measured confounders an unlikely sole explanation for model performance, although the cross-sectional case-control design and possible residual confounding preclude causal or prediagnostic claims. Together with the low prevalence-adjusted PPV and high NPV, these findings position retinal analysis as a potential triage or rule-out tool rather than a stand-alone diagnostic classifier. Conclusion: DL analysis of retinal fundus photographs can discriminate CRC status with high accuracy in a case-control setting, supporting the presence of CRC-associated retinal signatures. Prevalence-adjusted projections indicate limited stand-alone screening utility, positioning retinal AI analysis primarily as a potential triage or risk-enrichment tool pending longitudinal validation.