Alex Zhu alexzzhu

I received my PhD from the University of Pennsylvania working on algorithms for event based cameras. I am now a Senior Research Scientist at Waymo.

alexzzhu / compress_video

Last active February 24, 2018 20:08

Compresses a video using ffmpeg for conference submissions

	# First, resize the video to a reasonable size. Here only the width is specified, and the height is determined by the
	# aspect ratio.
	ffmpeg -i input.mp4 -vf scale=1024:-2 -an input_resized.mp4

	# Here a bitrate of 300k is used. To calculate the right bitrate, use the formula:
	# bitrate = FILE_SIZE81921000/LENGTH
	# where FILE_SIZE is in MB and LENGTH is in seconds, and the output bitrate is in kilobits.
	# e.g. for a 137s video and a desired file size of 5MB, the bitrate should be around 581921000/137 ~ 300k
	ffmpeg -y -i input_resized.mp4 -c:v libx264 -preset medium -b:v 300k -pass 1 -an -f mp4 /dev/null
	ffmpeg -i input_resized.mp4 -c:v libx264 -preset medium -b:v 300k -pass 2 -an input_compressed.mp4

alexzzhu / kalibr_to_ros_cinfo.py

Created March 20, 2018 17:40

Python script to generate ROS format yaml files from the output of Kalibr. Also useful to compute rectification and projection matrices with OpenCV.

	#!/usr/bin/env python
	import numpy as np
	import cv2
	import argparse
	import ruamel.yaml

	def main():
	parser = argparse.ArgumentParser(description="Converts a cam info in kalibr yaml format to the expected ROS format for camera_info_manager.")
	parser.add_argument("--camchain", dest="camchain",
	help="Camchain yaml file with camera info.",

alexzzhu / create_gif.sh

Last active November 27, 2018 18:52

Creates a gif from a sequence of images

convert -delay 1x60 -loop 0 *.png video.gif