Defendish- No Room for Error, No Reason for Fear

Abstract

Food allergies represent a growing public health concern in India, particularly among pediatric populations. The complexity of Indian food labeling, combined with the presence of multilingual packaging and a lack of standardized allergen warnings, creates a high-risk environment for allergic individuals. This paper presents Defendish, a comprehensive mobile health (mHealth) system designed to eliminate human error in allergen identification. Utilizing Flutter for cross-platform accessibility, Google Cloud Vision AI for Optical Character Recognition (OCR), and a hybrid Firebase-PostgreSQL backend, Defendish provides real-time ingredient analysis and safe/unsafe decision logic. The system introduces a community-driven data enrichment model tailored to the Indian market. Preliminary evaluations indicate high accuracy in allergen detection and a significant reduction in parental anxiety levels. Food allergies pose a serious and growing health concern, particularly among children, where accidental exposure to allergens can lead to severe and life-threatening reactions. In India, this challenge is intensified by complex food labeling practices, multilingual ingredient lists, and the widespread consumption of unpackaged or locally produced food items. As a result, parents and caregivers experience persistent anxiety while making everyday food choices, as even a minor oversight can have critical consequences. This paper presents Defendish, an AI-powered mobile application designed to provide real-time allergen detection and enhance food safety for Indian families. Defendish transforms a smartphone into an intelligent food safety assistant by allowing users to scan barcodes or ingredient labels using the device camera. The system analyzes food composition by combining cloud-based optical character recognition with a structured allergen database and personalized user allergy profiles. The application delivers immediate and unambiguous feedback by classifying food items as safe or unsafe based on detected allergens. Built using Flutter for cross- platform compatibility, Firebase for scalable backend services, PostgreSQL for structured data management, and Google Cloud Vision AI for ingredient recognition, the system ensures accuracy, responsiveness, and scalability. Additionally, Defendish incorporates a community-driven data model to continuously expand its food database, making it adaptable to India’s diverse food ecosystem. By prioritizing safety, simplicity, and accessibility, Defendish aims to reduce food- related anxiety and empower families to make confident dietary decisions, thereby contributing to improved public health outcomes.

Introduction

Food allergies in India are rising due to changing diets and increased processed food consumption, creating daily anxiety for parents of children with severe allergies. Inconsistent labeling and complex ingredient names make manual allergen management unreliable. Although FSSAI mandates allergen labeling, enforcement is weak, and local foods are often unlisted in global allergen databases.

Problem: Existing methods rely on manual label reading, which is prone to errors due to complex ingredients, language barriers, data gaps for Indian brands, and restaurant miscommunications.

Proposed Solution – Defendish:
Defendish is a mobile ecosystem designed to provide real-time, automated allergen identification for Indian families, based on a “Zero-Trust” safety model: unknown or unrecognized ingredients trigger a warning rather than assuming safety.

Key Features:

• Instant Identification: AI-based real-time processing of labels and barcodes.

• Multi-Profile Management: Supports multiple family members’ allergy profiles.

• Digital Allergy Card: Bilingual card (English/Hindi) for restaurants or social settings.

• Crowd-Sourced Database: Community enrichment for local and unbranded items.

System Architecture & Technology:

• Frontend: Flutter for cross-platform performance.

• Backend: Firebase for authentication and real-time triggers.

• Database: PostgreSQL for allergen hierarchies and product metadata.

• AI Engine: Google Cloud Vision OCR for text extraction from packaging.

Methodology:

1. Users create personalized allergy profiles.

2. Food data is acquired via barcode scans (packaged items) or OCR-based image scans (unpackaged/local foods).

3. Text is processed, normalized, and compared against allergy profiles using a conservative, safety-first algorithm.

4. Results are displayed clearly, and anonymized data improves the system over time.

Results:

• Accuracy: 94% in identifying allergens in 100 Indian packaged goods.

• Latency: Average scan-to-result time: 1.8 seconds on 4G.

• User Impact: 85% of parents in a pilot reported reduced stress and faster shopping.

• OCR struggles with handwritten labels, especially on local snacks.

Social Impact:
Defendish empowers caregivers, reduces anxiety, and improves safety for children with allergies, addressing a critical gap in India’s public health system by enabling proactive prevention rather than reactive treatment.

Conclusion

Defendish provides a robust, AI-driven solution to the growing problem of food allergies in India. By integrating OCR technology with a localized database and a user- centric design, the system effectively removes \"room for error\" in ingredient screening. As the database grows and AI models refine, Defendish aims to become the gold standard for food safety management in Indian households.

References

[1] Sicherer, S. H., “Food allergy management and prevention strategies,” Journal of Allergy and Clinical Immunology, vol. 135, no. 5, pp. 1186–1197, 2015. [2] Gupta, R. S., et al., “Food allergy awareness and safety challenges among children,” Pediatrics, vol. 140, no. 2, pp. 1–10, 2017. [3] World Health Organization, Food Safety and Consumer Health, WHO Press, Geneva, 2021. [4] Turner, P. J., & Campbell, D. E., “Severe food allergies and risk reduction approaches,” Clinical and Experimental Allergy, vol. 46, no. 3, pp. 327–336, 2016 [5] Free, C., Phillips, G., & Watson, L., “Mobile health technologies and their role in public health systems,” Journal of Medical Internet Research, vol. 15, no. 2,pp. e56, 2013. [6] Islam, S. M. S., et al., “Mobile applications for healthcare decision support,” JMIR mHealth and uHealth, vol. 3, no. 3, e62, 2015. [7] Baviskar, P., & Kulkarni, S., “Cross-platform mobile application development using Flutter,” International Journal of Computer Applications, vol. 182, no. 44, pp. 1–6, 2018.

Copyright

Copyright © 2025 Dr. Monika Bhatnagar, Abhijeet Kumar, Anjesh Kumar, Himanshu Kumar, Vaibhav Bharadwaj. This 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.