Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{191561,
        author = {Mr. Uttam Tiwari and Ms. Neha Kumari},
        title = {MooNet: A Lightweight CNN for Mobile Cattle Breed Recognition with QR-Based Digital Traceability},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {7712-7721},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=191561},
        abstract = {The correct identification of cattle breeds is a factor related to livestock management, genetic assessment, and the maintenance of trustworthy electronic data. Practically, non-expert operators tend to mislabel morphologically similar indigenous and crossbred animals where images are taken at varying lighting, motion blur, occlusions and with a crowded farm background. This compromises the quality of big livestock databases. This paper introduces MooNet, a mobile and edge-delivered convolutional neural network that is based on MobileNetV4 and capable of recognizing cattle by their breed in real-time. The 12,000 annotated images used to train MooNet are field-based data as they have inherent variation of the viewpoint, light, and image complexity. The model performs at 96.45% accuracy, 96.47% precision, 96.45% recall, and 96.45% F1-score on held- out validation split, which shows even-handed performance across classes. It is a depth wise separable architecture resulting in low-latency inference with an average mean performance of about 17.95 MS per image on mid-range mobile platforms, which makes it appropriate in resource-constrained deployments. All animals in the categories can have a digital identity based on QR, which holds breed and history information, making it a tool to track them over time. The proposed system will allow decreasing the number of errors in manual entries and enhance the quality of information systems on livestock because it is implemented using high-precision visual recognition with persistent per-animal profiles.},
        keywords = {Recognition in cattle breeding, lightweight CNN, MooNet architecture, livestock identification on mobile phones, traceability via QR-codes, field-level image-based cattle identification.},
        month = {January},
        }