csrnet

(ECCV 2020) Conditional Sequential Modulation for Efficient Global Image Retouching

Implementation of the CSRNet paper (CVPR 18) in keras-tensorflow

SOFT-CSRNET : Counting people in drone video footage

An unofficial implementation of CSRNet for crowd counting.

Real-time crowd density estimation & interactive flow simulation powered by CSRNet. Trained on ShanghaiTech and custom Indian crowd datasets with FastAPI & Vite.

Train a CSRNet for estimating a crowd distribution

Implement CSRnet for counting vehicle traffic

A deep learning project implementing CSRNet for accurate people counting in highly congested scenes, based on the architecture described in the CSRNet paper.

Keras based implementation of the CSRNET model

Real-time crowd intelligence platform — YOLOv8 + CSRNet + multi-horizon LSTM risk forecasting

AI-powered smart city crowd management and dynamic bus dispatch system using CSRNet density map regression to monitor transit stop overcrowding and optimize public transportation.

Crowd counting and behavior analysis web app using PyTorch (CSRNet) with a Flask-based frontend.

Dense crowd counting with CSRNet + CBAM attention — MAE 61.14 / RMSE 90.91 on ShanghaiTech Part A, beating the CVPR'18 CSRNet baseline (68.2 / 115.0) — paired with a fully local Ollama multimodal agent that explains every prediction.

Repo for computer vision dataset of trees around the London area

Real-time crowd density estimation using CSRNet (VGG-16 + dilated CNN) on ShanghaiTech Part A & B. Includes adaptive Gaussian kernel density map generation, from-scratch training on M1 CPU, MPS inference, and a live video demo with heatmap overlay and running count graph. Part A MAE: 87.85 | Part B MAE: 21.12

CPU-efficient crowd counting: a reproducible study of how INT8 quantization and input resolution trade off accuracy against latency, using CSRNet and ONNX Runtime

PyTorch training and evaluation framework for crowd counting density estimation models (CSRNet, resSFCN, MCNN) on standard and custom datasets.

A Django web application providing real-time crowd counting and density heatmap overlay streaming using pre-trained CSRNet models.

Single-Image Crowd Counting - Computer Vision Project