http://statkclee.github.io/think-stat/

통계적 사고 워크샵

싸이그래머와 xwMOOC가 함께 합니다.

기계와의 경쟁을 준비하며…

“The future is here, it’s just not evenly distributed yet.”
- William Gibson

학습 개요

• 일시
  • ’15년 11월 18일 ~ ’16년 1월 6일 (8주간)
  • 매주 수요일 19:00 ~ 20:00
• 장소
  • 서울 강남역 근처 토즈 타워점 : 위치확인

교재

통계적 사고 (ThinkStats2) : 프로그래머를 위한 확률과 통계
- (번역): https://github.com/statkclee/ThinkStats2
- (원서): http://greenteapress.com/thinkstats2/
- 번역 버젼(’15년 11월 17일), PDF 파일
- 번역 버젼(’15년 11월 17일), HTML 파일
«««« 연습문제 초벌 번역
«««« 한국어판 서문 추가

장연습문제해답

공용 작업공간

위키…통계적 사고

LaTeX 사전 공부

• LaTeX 기초
  • LaTeX 소개 슬라이드
  • LaTeX 소개 문서형식
• LaTeX 석박사 논문 저작
  • LaTeX : 수학, BibTeX, 사용자 정의

출처: OpenIntro.org LaTeX 저작 교육과정

사전 준비

1. Git과 GitHub
   • 소프트웨어 카펜트리 Git을 이용한 버젼관리
   • 리눅스와 맥 사용자: sudo apt-get install git
   • 윈도우 사용자: Git Bash 설치
2. 파이썬 설치: 아나콘다 기반 과학컴퓨팅 개발환경을 적극추천
   • 아나콘다 기반 설치
   • 파이썬 기반 설치
   • 아나콘다를 적극 추천하는 사유

# 파이썬 2 기준 우분투 14.04 64비트 버젼 설치 사례 $ wget https://3230d63b5fc54e62148e-c95ac804525aac4b6dba79b00b39d1d3.ssl.cf1.rackcdn.com/Anaconda2-2.4.0-Linux-x86_64.sh $ bash ~/Downloads/Anaconda2-2.4.0-Linux-x86_64.sh 사전준비 검증[xwmooc:~/ThinkStats2/code ] $ python nsfg.pynsfg.py:42: SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy df.birthwgt_lb[df.birthwgt_lb > 20] = np.nan (13593, 244) nsfg.py: All tests passed.부랑자(Vagrant)와 가상상자(VirtualBox)를 활용한 과학컴퓨팅 환경 구축

부랑자(Vagrant) 아나콘다 설치환경 참조

학습 목차

일시학습 내용

열린 통계학 개론

• 현재 작업 중 :열린 통계학 개론

언론 기사

• 한국통계진흥원에서 발간 통계의 창 (2015년 여름호)
  • 통계교육원 > 열린교육방 > e-book 게시판
  • 미래인재 데이터과학지: 교육사례중심
  • 전체 다운로드(100MB)
• 강서양천신문 ’15년 9월
  • 기고, 컴퓨터 사고력과 소프트웨어 교육
  • 기고, 글을 쓰는 것은 소프트웨어 개발이다!

참고자료

• $100 달러 오픈 컴퓨터 : $100 달러 오픈 컴퓨터 워크샵
• 컴퓨터과학 언플러그드
• 러플(Rur-ple)
• 파이썬 거북이
  • 소프트웨어 카펜트리 거북이 (영문)
  • 소프트웨어 카펜트리 거북이 (국문)
• 정보교육을 위한 파이썬
• 라즈베리 파이
• 소프트웨어 카펜트리
• R 팩키지

xwMOOC 한글소스 Contact License

https://github.com/orlyliba/face-masker

Code for automatically putting a mask on a face in photographs. The code matches features between the mask and the face in the photo and warps the mask and merges it onto the face. Additionally, the code adds ears or a hat onto the photo.

face-masker

Code (Matlab) for automatically putting a mask on a face in photographs. The code matches features between the mask and the face in the photo and warps the mask and merges it onto the face. Additionally, the code adds ears or a hat onto the photo.

In order to detect matching feature points in the faces and masks, a separate algorithm should be used (this repository does not include it). I have tested two feature detectors that worked well:

1. Face++ (http://www.faceplusplus.com/), in which feature detection is done in the cloud
2. Clandmark (http://cmp.felk.cvut.cz/~uricamic/clandmark/), an open source landmarking library

Masks should be stored as the basis image, feature points and blending mask (binary or gray scale weights). The masks can also contain ears/hat/wig. The feature points are detected by the same algorithm that you plan on using for photos, and can be shifted manually. The blending masks are best when done manually with simple programs such as paint.

Read the report and look at the poster to learn more about this project.

[1] document: https://pychubby.readthedocs.io/en/latest/source/basic_example.html

[2] 소스코드: https://github.com/jankrepl/pychubby

[3] 블로그: https://jankrepl.github.io/pychubby/

----------------------------------------------------------------------------------------------------

[1]

Basic Example

To illustrate the simplest use case let us assume that we start with a photo with a single face in it.

pychubby implements a class LandmarkFace which stores all relevant data that enable face warping. Namely it is the image itself and 68 landmark points. To instantiate a LandmarkFace one needs to use a utility class method estimate.

import matplotlib.pyplot as plt from pychubby.detect import LandmarkFace img = plt.imread("path/to/the/image") lf = LandmarkFace.estimate(img) lf.plot()

Note that it might be necessary to upsample the image before the estimation. For convenience the estimate method has an optional parameter n_upsamples.

Once the landmark points are estimated we can move on with performing actions on the face. Let’s try to make the person smile:

from pychubby.actions import Smile a = Smile(scale=0.2) new_lf, df = a.perform(lf) # lf defined above new_lf.plot(show_landmarks=False)

There are 2 important things to note. Firstly the new_lf now contains both the warped version of the original image as well as the transformed landmark points. Secondly, the perform method also returns a df which is an instance of pychubby.base.DisplacementField and represents the pixel by pixel transformation between the old and the new (smiling) image.

To see all currently available actions go to Gallery.

To create an animation of the action we can use the visualization module.

from pychubby.visualization import create_animation ani = create_animation(df, img) # the displacement field and the original image

------------------------------------------------------------------------------------------------------

[2]

PyChubby

Tool for automated face warping

Installation

pip install pychubby

If you get an error FileNotFoundError: [Errno 2] No such file or directory: 'cmake': 'cmake', you need to make sure cmake is installed. If you're on OSX you can install this via Homebrew with:

brew install cmake

For other platforms please consult the Cmake documentation at https://cmake.org/install/

Description

For each face in an image define what actions are to be performed on it, pychubby will do the rest.

Documentation

https://pychubby.readthedocs.io

Minimal Example

import matplotlib.pyplot as plt from pychubby.actions import Chubbify, Multiple, Pipeline, Smile from pychubby.detect import LandmarkFace img_path = 'path/to/your/image' img = plt.imread(img_path) lf = LandmarkFace.estimate(img) a_per_face = Pipeline([Chubbify(), Smile()]) a_all = Multiple(a_per_face) new_lf, _ = a_all.perform(lf) new_lf.plot(show_landmarks=False, show_numbers=False)

CLI

pychubby also comes with a CLI that exposes some of its functionality. You can list the commands with pc --help:

Usage: pc [OPTIONS] COMMAND [ARGS]... Automated face warping tool. Options: --help Show this message and exit. Commands: list List available actions. perform Take an action.

To perform an action (Smile in the example below) and plot the result on the screen

pc perform Smile INPUT_IMG_PATH

or if you want to create a new image and save it

pc perform Smile INPUT_IMG_PATH OUTPUT_IMG_PATH

Development

git clone https://github.com/jankrepl/pychubby.git cd pychubby pip install -e .[dev]

------------------------------------------------------------------------------------------------------

[3]

PyChubby - Automated Face Warping

 2 minute read

Introduction

We all know the story. You ask a random person on a street to take a photo of you and your friends. After a few moments you happily thank them and go on with your life. After some time you finally sit down for a cup of coffee and check your photos. “My god, why is none of us smiling?!”.

First of all, do not panic. It is your lucky day. If you know how to pip install things there might be hope.

pip install pychubby

Once installed, just write something like this:

import matplotlib.pyplot as plt from pychubby.actions import Multiple, Smile from pychubby.detect import LandmarkFace img_path = 'path/to/img.jpg' img = plt.imread(img_path) lf = LandmarkFace.estimate(img) a = Multiple(Smile(0.15)) lf_new, _ = a.perform(lf) lf_new.plot(show_numbers=False)

What is PyChubby

Pychubby is an automated face warping tool. Its main goal is to serve as a specialized augmentation interface for deep learning face related tasks. But it might as well be used as a silly face warping tool (see the introduction).

You might wonder why even bother when one can do similar things with Photoshop and other software. The answer is simple - automation. You do not have to locate any landmarks, move them around and then repeat the procedure on each face in every photo.

Popular image augmentation packages like imgaug are general purpose (any kind of image) and do not provide many options when it comes to geometric transformations. pychubby is specialized on human faces and allows for creation of geometric transformations that are:

1. Local (on the face)
2. Smooth (no artifacts)
3. Realistic (to an extent)

In other words one does not augment the whole image but just the faces. The augmentations are realistic and have no artifacts.

Building blocks

The logic of pychubby can be summarized in three blocks

1. Landmark Detection
   • Given a photo, a pretrained landmark detection model predicts 68 landmarks on each face.
2. Reference Space Mapping
   • The landmarks are mapped into a so called reference space. This mapping corrects for possible rotations, translations and scaling in the input image.
3. Manual Action Definition
   • Majority of pychubby actions are defined in the reference space and therefore should yield consistent warpings across different faces. Go to Gallery to see the predefined ones or feel free to define new ones.

Want to know more?

If you are interested in giving pychubby a try or just want to learn more see below a few useful links:

• GitHub
• Documentation

All potential contributors are more than welcome and any feedback is highly appreciated.

------------------------------------------------------------------------------------------------------