subarnatripathi Goto Github PK

arbitrary_resolution_change

This can change the resolution of any 420 yuv image to any arbitrary user input resolution. Among the three input algorithms supported (nearest neighbor, bilinear filter and polyphase filter), the polyphase filter implementation reuses a significant code from ffmpeg library. I only simplified some parts.

densecrf_multiclass_video_segmentation

Dense CRF for multiclass video labelling

gptk

Automatically exported from code.google.com/p/gptk

motion_segmentation

pre-alpha video motion segmentation code

object-tracking

This OpenCv based C implementation is for tracking multiple ROI within a video. This is a an appearance based tracker exploiting particle filter and importance sampling.

reversegan

Recovering latent vectors from Generative Adversarial Networks

rnn_video_object_detection

sds_eccv2014

streamgbhpp

Improving StreamGBH with motion

streaming_vop_clustering

subarnatripathi.github.io

tracker_initialization

This matlab code is for Tracker initialization.

video_inference

mean field based inference for dense semantic video segmentation with higher order clique potentials (Pn Potts model)

yolotf

translate any darknet config to tensorflow, including YOLO. Load trained weights, retrain/fine-tune them using tensorflow, export to protobuf file for iOS dev