Real-Time Object Detection with Audio Feedback

[Shraddha-DSA]

Hello @kaushav07 and team 👋,

I’d like to work on implementing real-time object detection with audio feedback as part of VisionMate’s assistive features.

🔧 Plan:

Use YOLOv8 for detecting objects from the webcam feed

Use pyttsx3 for converting detections to spoken audio

Overlay bounding boxes and labels using OpenCV

Modular codebase in utils/object_tts.py

📁 Folder structure:
VisionMate/
└── utils/
└── object_tts.py
Let me know if this looks good — I’ll start working and raise a PR!

Thanks!
Shraddha (@Shraddha-DSA)

[omgupta-JUS]

Hi @kaushav07 👋,

I'm Om Gupta, a GSSoC’25 contributor and an AI full stack developer in training. I’m genuinely inspired by the mission behind VisionMate, and I’d love to contribute to the Real-Time Object Detection with Audio Feedback module.

🧠 Proposed Plan:
Implement an optimized YOLOv5/YOLOv8 or MobileNet SSD model for real-time object detection using OpenCV + Python

Integrate Text-to-Speech (using pyttsx3 or gTTS) to deliver immediate audio feedback for detected objects

Build a modular pipeline that allows easy scaling for future obstacle detection or environment classification features

Optionally provide a simple CLI or Flask-based demo interface for local testing

🔧 Tech I'll Use:
Python, OpenCV

PyTorch or TensorFlow (for pretrained object detection)

pyttsx3 or gTTS for audio feedback

Modular codebase suitable for future integration with frontend (React or Flutter)

I’d be thrilled if you could assign this issue to me. Looking forward to contributing meaningfully to this impactful project! 🔊👁️‍🗨️

Thanks & regards,
Om Gupta

[shreeraksha2112]

@Shraddha-DSA This issue is assigned to you. Please only raise those issues which haven't been raised or the ones you feel that it needs to be improved. Would be glad to help you further.

[Umisalma848]

Hi @kaushav07,

I hope you're doing well! I came across the "Real-Time Object Detection with Audio Feedback" issue and found it really interesting and aligned with my skills. I would love to contribute to this feature.

Here are a few ideas I'd like to explore while working on it:

1. Use a pre-trained YOLOv5 or YOLOv8 model for real-time object detection via webcam.
2. Integrate pyttsx3 or gTTS for generating audio feedback based on detected objects.
3. Add a feature to limit audio output to unique or important detections to avoid repetition.
4. Optional GUI using Tkinter or Streamlit to show detections visually along with the audio.

I’d be really grateful if you could kindly assign this issue to me. Looking forward to your feedback!

Thanks and regards,
Umi Salma
GSSoC '25 Contributor

[shreeraksha2112]

@Umisalma848 This issue has already been assigned to two of them. You can collaborate and discuss with them or raise another issue of your own.