emv aka emphatetic visual assistant for blind people

mv empehatetic visual assistant is a sotware that helps blind peop;e to walk with confidence

1. Empathetic Vision Assistant (EVA) Team Members: Sarvesh Shinde (462) Keyur Shinge (463) Prachit Yelamkar (470) Parth Rapase (448) Guide: Dr. Nilesh Bhelkar

2. ABSTRACT Empathetic Vision Assistant (EVA) is an AI-powered system designed to assist visually impaired individuals by combining real-time object detection (YOLO) and emotion recognition models. EVA provides empathetic guidance by interpreting the surrounding environment, recognizing obstacles, and detecting human emotions to improve safety and social awareness. The backend is implemented using Django REST API with Python models, while the frontend is built in Flutter for a mobile-first experience. EVA transforms AI into an inclusive assistant by offering multilingual support, real-time feedback, and a seamless integration of computer vision and emotional intelligence.

3. INTRODUCTION Welcome to EVA, an innovative project aimed at enhancing accessibility for visually impaired individuals. EVA leverages computer vision, deep learning, and emotional intelligence to provide real-time assistance. Through YOLO object detection and emotion recognition, EVA alerts users about obstacles and interprets social context. The system uses a Django backend for processing and a Flutter frontend for seamless user interaction. By providing empathetic assistance, EVA ensures safer navigation and improved independence.

4. LITERATURE SURVEY 1. J. Redmon et al., YOLO: Real-Time Object Detection, CVPR 2016 – efficient object detection framework. 2. K. Simonyan, A. Zisserman, VGGNet (2014) – CNN architectures for deep feature extraction. 3. MediaPipe & TensorFlow Lite – lightweight frameworks for real-time face and emotion detection. Research Gap: Most assistive systems focus only on obstacle detection and ignore emotional/social awareness. EVA bridges this gap by combining object detection with emotion recognition for empathetic AI assistance.

5. PROBLEM STATEMENT Visually impaired individuals face critical challenges such as: - Difficulty in navigating crowded or unfamiliar environments - Risk of collisions with obstacles - Lack of awareness about people's emotions or intentions Existing systems provide navigation assistance but lack empathetic intelligence to interpret social context.

6. OBJECTIVES 1. Develop real-time object detection using YOLO. 2. Integrate facial emotion recognition for empathetic guidance. 3. Provide mobile-first accessibility using Flutter frontend. 4. Implement Django REST API backend for model integration. 5. Enable multilingual interaction for inclusivity. 6. Ensure independence and safety for visually impaired individuals.

7. EXISTING SYSTEM - Traditional assistive technologies focus mainly on obstacle detection. - Navigation systems provide guidance but lack emotional/social awareness. - Systems such as smart canes or basic vision assistants are limited to pathfinding. Limitations: - No integration of emotion recognition - Limited real-time processing - Lack of multilingual and empathetic interaction

8. PROPOSED SYSTEM EVA integrates multiple technologies: - YOLO for object detection - Emotion detection model for analyzing social context - Django REST API backend for ML processing - Flutter frontend for real-time user interaction System Flow: 1. Capture input via Flutter app 2. Send to Django backend 3. Object + Emotion detection 4. JSON Response 5. Convert to Audio feedback [Placeholder for System Architecture Diagram]

9. IMPLEMENTATION - Backend: Python, Django REST Framework - Models: YOLO for object detection, CNN/MediaPipe for emotion recognition - Frontend: Flutter app with TTS Outputs: - JSON responses with objects + emotions - Voice-based feedback for visually impaired users [Placeholder for Screenshots and Results]

10. REFERENCES [1] J. Redmon et al., YOLO: Real-Time Object Detection, CVPR 2016 [2] K. Simonyan, A. Zisserman, VGGNet, arXiv, 2014 [3] TensorFlow Documentation: https://www.tensorflow.org [4] Flutter Documentation: https://flutter.dev [5] Django REST Framework: https://www.django-rest-framework.org

11. THANK YOU