Object detection using Faster rcnn

Shrimp Seed Counter Project
I developed an AI-powered Shrimp Seed Counter system to automate the counting process of shrimp seeds in aquaculture farms. Traditionally, shrimp seeds are counted manually using spoons or small containers, which often leads to inaccurate counts, financial loss, and inconsistency for both farmers and hatchery providers.
To solve this problem, I built a computer vision-based solution using Deep Learning and Object Detection techniques. The system takes live camera images or uploaded images as input and automatically detects and counts shrimp seeds with high accuracy.
Key Features
Real-time shrimp seed detection and counting
AI-based object detection using Faster R-CNN / Detectron2
Live camera support
Annotated output images showing detected shrimp seeds
Automatic count generation
User-friendly interface built with Streamlit/Gradio
Technologies Used
Python
PyTorch
Detectron2 / Faster R-CNN
OpenCV
Roboflow for dataset management
COCO annotation format
Streamlit/Gradio for frontend deployment
Workflow
Collect shrimp seed images from aquaculture environments
Annotate shrimp seeds using bounding boxes
Train an object detection model on the dataset
Detect shrimp seeds from new images/video streams
Generate accurate counts and annotated outputs
Impact
This project helps:
Reduce manual counting effort
Improve counting accuracy
Save time for aquaculture farmers
Minimize financial losses caused by inaccurate seed estimation
The project demonstrates practical application of AI and computer vision in the aquaculture industry and showcases skills in deep learning, model training, deployment, and real-world problem solving.

Видео Object detection using Faster rcnn канала Yaseen Mohammad