Temporally Identity-Aware SSD with Attentional LSTM

detect objects from cctv camera

영상 등급 분류 보조 시스템

A Dual Refinement Mechanism for Real-World Visual Detection

This python code uses OpenCV to detect movements in videos and logs with timestamps. Special features: 1. set region of interest by coordinates 2.set noise level 3. set how clean the log you want.

Object detection with ssd_mobilenet and tiny-yolo (Add: YOLOv3, tflite)

detect the no of people every second entering building gate. #person-detection

Pistol, Rifle, and Fire detection using yolov4-tiny in videos as well as images. Training code, dataset, and trained weight file available.

This project combines YOLOv2 and seq-nms to realise real time video detection. Main contribution: creation of easy to follow insctructions.

mask rcnn training with coco-like dataset. You can use for trainnig your own coco.json (polygon) dataset in Google Colab.

Implementation of the seq-nms post-processing algorithm for video object detection

In this project I tried to make a program that detects objects in videos using the open cv library and the Yolo dataset .

Smartbin AI is a personal project aiming at training an artifial intelligence to recognise waste material using Tensorflow.

Real Time Detection of Anomalous Activity From Videos (mainly crime actvity). Images of the video is trained using AutoEncoder to get the imtermediate feature representation of image & applied svm model for the bag of such features to detect the anomaly & LSTM to detect the type of Anomaly.

Road Lane Dection using OpenCV in python

Software y tecnología aérea, espacial y cibernética para la protección de la Amazonía - Ganadores del Codefest 2023

Authors official PyTorch implementation of the "Self-Supervised Video Similarity Learning" [CVPRW 2023]

YOLO V8 video detection inside TouchDesigner

Repository for the paper "Can lies be faked? Comparing low-stakes and high-stakes deception video datasets from a Machine Learning perspective"

视频相似度检测（vidsimdtion）是由PlugLink官方开发自媒体视频相似度检测，您或许会遇到在矩阵号发布种草视频容易被限流和下架，这是因为视频过于相似造成的，该检测工具可以让你清楚每一批视频的相似度，进而方便你决定哪些要发哪些不要发，一般相似度低于70%的视频发布成功率较高，一些不是很严格的平台可能在80%以下就能发了。