• audio_landmarks3d: 3d face landmarks fitted to speech audio
• blink.cpp: eye-blink classification based on facial landmarks
• bow: BOWKmajority example(for binary descriptors) also, bow with python in opencv 3.x
• chili: opencv c++ remote playground running on heroku
• cluster: clustering att faces with kmeans,flann,random and BRUTEFORCE!!
• curl.cpp: download a car from the internet ;)
• colab: some jupyter notebooks, mostly tf/pytorch -> dnn conversions
• condens2.cpp: 2.x adaption of the old legacy condensation code
• db: serialize cv::Mat to some popular databases.
• dbrief: c++, opencv2.x port of http://cvlab.epfl.ch/research/detect/dbrief
• demo: various c++/py samples, mostly related to problems found on answers.opencv.org or #opencv
• dnn_dot: graphviz visualization of opencv's dnn models
• eyes.cpp: hog/svm based open / closed eyes detector, using flandmarks for eye location
• facerec_dnn: using opencv's dnn module and a pretrained (torch) facenet for transfer learning. also a realtime webcam demo.
• facerec_online.py: realtime facerec from webcam ( with adaptive face-db )
• facerec_online.cpp: same in c++
• gas.cpp: Neural Gas (clustering) experimentBRUTE FORCE!!!!
• gldemo.cpp : opengl inside opencv (+ minimal .obj reader)
• java_dnn: java examples for dnn models
• java_face3.0: testcase for ocv3 facerec bindings in java
• headpose: small head pose estimation (using dlib & pico)
• hep.cpp: playing with http://dismac.dii.unipg.it/hep/index.html
• html5: facedetect from webcam in browser ( using liu liu's js code )
• ica: An Introduction to Independent Component Analysis: InfoMax and FastICA algorithms
• kalman.cpp: mouse-tracking kalman demo
• kotlin: jikes, opencv java bindings in kotlin ;)
• mace: Minimum Average Correlation Energy Filter, useful for authentification with (cancellable) biometrical features from small datasets.
• mjpg_serve.py: mjpeg webcam server ( single-thread, single-client, sorry )
• mjpg_serve.cpp: mjpeg webcam server ( single-thread, but multi client )
• ml/libsvm_ocv.cpp : using libsvm with opencv
• openpose_multi.cpp: an attempt to use opencv's dnn with multiple persons, exploiting the PAF's in the original openpose model (WIP)
• profile.h: profiling / timing code
• pushbroom: an adaption of pushbroom-stereo https://github.com/andybarry/flight
• shape_match: some shape/contours/gestures playground (various distances with a common interface).
• skin.cpp: skin detector (Rehg/Jones,1999)
• tcnn_landmarks: caffe_dnn, a small example for pretrained facial landmarks from tal hassner's tcnn project
• tsc: trafic sign classification with https://github.com/tiny-dnn/
• toy/echo: echo chamber
• toy/zoom: zooming in/out of highgui window
• toy/kaleidoscope: .. kids ..
• tracking: BACF and CMT tracker
• ui : onscreen ui (sliders & buttons & selectboxes)
• vibe.cpp: ViBe background separation
• violentFlow.cpp : http://www.openu.ac.il/home/hassner/data/violentflows/violent_flows.pdf