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{206065,
        author = {Jyoti and Swati Khanve},
        title = {Mapping the Multi-Dimensional: A Review of Deep Feature Representation and Similarity Search},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {2},
        pages = {1562-1565},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=206065},
        abstract = {Driven by the evolution of deep convolutional neural networks (CNNs), modern Content-Based Image Retrieval (CBIR) has shifted from low-level visual parsing to high-dimensional semantic embedding and scalable proximity architectures. While foundational CBIR paradigms rely on handcrafted descriptors (e.g., SIFT, GLCM, and color histograms), they inherently suffer from the "semantic gap"—failing to capture abstract contextual data and limiting retrieval precision. This survey comprehensively investigates ten pivotal studies spanning 2020–2025, with a distinct emphasis on landmark contributions from Indian researchers. We dissect the trajectory of deep feature extraction, hybrid retrieval pipelines, multi-dimensional indexing frameworks, and algorithmic optimizations. Furthermore, this paper identifies critical bottlenecks in current literature and introduces a consolidated, generalized framework that synergizes CNN-driven feature extraction, Principal Component Analysis (PCA) dimensionality reduction, and Approximate Nearest Neighbor (ANN) indexing variants like FAISS and HNSW to optimize the accuracy-scalability trade-off.},
        keywords = {Content-Based Image Retrieval (CBIR), High-Dimensional Embeddings, Semantic Gap, Deep Feature Extraction.},
        month = {July},
        }