CITATION.cff

cff-version: 1.2.0
title: 'StockSight AI: Enterprise-Grade Dense Object Detection System'
message: >-
  If you use StockSight AI in your research, production systems,
  or academic work, please cite it using the metadata below.
type: software
authors:
  - given-names: Daksh
    family-names: Pathak
    email: daksh.pathak.ug24@nsut.ac.in
    affiliation: Independent Researcher
repository-code: >-
  https://github.com/dakshhhhh16/StockSite-AI
abstract: >-
  StockSight AI is an advanced, production-ready object detection framework
  optimized for dense retail environments. Built on YOLOv5 architecture and
  trained on the comprehensive SKU-110K dataset, this system achieves 92.2%
  mAP@50 while maintaining edge-device compatibility through sophisticated
  post-training quantization techniques. The framework supports multiple
  deployment formats (PyTorch, TensorFlow, TFLite) with INT8 quantization
  achieving 75% model size reduction with less than 1% accuracy degradation.
  StockSight AI enables real-time inference (30+ FPS) suitable for enterprise
  applications including inventory management, loss prevention, customer
  analytics, and supply chain optimization.
keywords:
  - Object Detection
  - Dense Object Detection
  - YOLOv5
  - SKU-110K
  - Computer Vision
  - Deep Learning
  - Model Quantization
  - INT8 Quantization
  - Edge Computing
  - TensorFlow Lite
  - Retail Analytics
  - Inventory Management
  - Real-time Inference
  - Enterprise AI
license: MIT
version: 2.0.0
date-released: '2025-11-23'