Open Source Deep Learning Serving System with Web Interface

基于 Pytorch 的垃圾识别与分类。本项目源于浙江大学光电学院课程设计。（程序仅供参考，很久没关注这方面内容了，为避免误导，相关问题不作回答）

A list of useful resources in the trash classification and detection (mainly plastic), such as datasets, papers, links to open source projects

Managing waste in fun and easy way with AI ♻️😊🤖

Ramudroid, autonomous solar-powered robot to clean roads, realtime object detection and webrtc based streaming

基于tensorFlowjs的垃圾分类

Python notebook about garbage detection based on convolutional neural network

A machine learning tool built with TensorFlow and the VGG16 model. It classifies waste items from images, assisting in efficient recycling. Users upload waste images, and the system identifies the waste type.

Official Droidrush repository of NPDevs team (Avishkar-2019--Annual Techfest of MNNIT Allahabad)

Welcome to the repository of our garbage classification project! We have developed a model using PyTorch and EfficientNet-B4 that classifies garbage into twelve different types. The model has achieved an impressive accuracy of 98.45%.

Metis project 6/7

This repository contains garbage classification models built using PyTorch, capable of accurately classifying garbage into different categories. It includes two versions of the model, one with 6 classes & the other with 12 classes, to cater to different needs. Change branch to "testbr" to see the 12-class model in the "Extra" directory.

Self-collected trash dataset used in AlphaTrash project. Contains 5600+ images of trash gathered in Thailand, sorted into general, metal, organic, paper, and plastic wastes.

This project implements a custom CNN model for waste classification (battery, plastic, etc.) into 12 categories. Trained on image data, it achieves 80.03% accuracy.

利用卷积神经网络进行垃圾分类识别

🥡一个开源的「垃圾分类」小程序。

FastAPI-based API system for ML garbage detection in 6 categories. Python-powered solution for efficient waste classification and management.

垃圾分类Android版应用, 拍照即可进行分析.

PeppeRecycle Android

A Pytorch project for garbage classification using the EfficientNet-B6 model to achive a 95.78% accuracy on the test set. 😊