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{198214,
        author = {Sonali and Mohd. Aman and Shadan Saifi},
        title = {Advanced Synergistic Framework for Age-Invariant Face Recognition: A Second-Version Integration of Deep Metric Learning and Generative Manifold Augmentation for Missing Person Identification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {14621-14629},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198214},
        abstract = {The identification of missing persons over protracted durations remains one of the most formidable challenges in the field of biometric forensics. The primary obstacle is the non-linear physiological transformation of facial structures over time, a process known as facial aging, which significantly degrades the performance of conventional face recognition algorithms. This research presents an exhaustive second-version framework for Age-Invariant Face Recognition (AIFR) that synergizes deep metric learning, generative adversarial manifolds, and high-performance vector indexing. At the core of the system is the Additive Angular Margin Loss (ArcFace), which enforces a clear geometric separation of identities on a hyperspherical embedding space. To bridge the temporal gap, we employ a Conditional Adversarial Autoencoder (CAAE) to synthetically populate the gallery with age-progressed variants of each subject, effectively transforming a sparse temporal dataset into a dense manifold. For real-time retrieval across million-scale databases, the framework utilizes Facebook AI Similarity Search (FAISS), specifically the Hierarchical Navigable Small World (HNSW) and Inverted File Product Quantization (IVFPQ) indices. Experimental results on the FG-NET and MORPH Album 2 datasets demonstrate a significant improvement in Rank-1 accuracy and F1-score compared to baseline models, with the proposed augmentation strategy yielding an absolute accuracy gain of approximately 3.33% in scenarios involving 40-year age gaps. This report provides a comprehensive mathematical derivation of the system components and offers deep insights into the optimization of cross-age biometric pipelines under the academic supervision of Shashikanth Maurya, whose expertise in neural networks and information retrieval grounds the technical robustness of this work.1},
        keywords = {},
        month = {April},
        }