1.功能描述:
對于輸入的一張圖片(可正常解碼,且長寬比適宜),檢測圖片中的所有人手,輸出每只手的坐標框、21個骨節點坐標資訊,
2.平臺接入
具體接入方式比較簡單,可以參考我的另一個帖子,這里就不重復了:
http://ai.baidu.com/forum/topic/show/943327
3.呼叫攻略(Python3)及評測
3.1首先認證授權:
在開始呼叫任何API之前需要先進行認證授權,具體的說明請參考:
http://ai.baidu.com/docs#/Auth/top
具體Python3代碼如下:
# -*- coding: utf-8 -*-
#!/usr/bin/env python
import urllib
import base64
import json
#client_id 為官網獲取的AK, client_secret 為官網獲取的SK
client_id =【百度云應用的AK】
client_secret =【百度云應用的SK】
#獲取token
def get_token():
host = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=' + client_id + '&client_secret=' + client_secret
request = urllib.request.Request(host)
request.add_header('Content-Type', 'application/json; charset=UTF-8')
response = urllib.request.urlopen(request)
token_content = response.read()
if token_content:
token_info = json.loads(token_content)
token_key = token_info['access_token']
return token_key
3.2手部關鍵點識別分析介面呼叫:
詳細說明請參考: https://ai.baidu.com/docs#/Body-API/2757b503
說明的比較清晰,這里就不重復了,
大家需要注意的是:
API訪問URL:https://aip.baidubce.com/rest/2.0/image-classify/v1/hand_analysis
影像資料,base64編碼后進行urlencode,要求base64編碼和urlencode后大小不超過4M,圖片的base64編碼是不包含圖片頭的,如(data:image/jpg;base64,),支持圖片格式:jpg、bmp、png,最短邊至少50px,最長邊最大4096px
Python3呼叫代碼如下:
#畫出手部識別結果
def draw_hands_point(originfilename,hands,resultfilename,pointsize,pointcolor):
from PIL import Image, ImageDraw
image_origin = Image.open(originfilename)
draw =ImageDraw.Draw(image_origin)
for hand in hands:
for hand_part in hand['hand_parts'].values():
#print(hand_part)
draw.ellipse((hand_part['x']-pointsize,hand_part['y']-pointsize,hand_part['x']+pointsize,hand_part['y']+pointsize),fill = pointcolor)
gesture = hand['location']
draw.rectangle((gesture['left'],gesture['top'],gesture['left']+gesture['width'],gesture['top']+gesture['height']),outline = "red")
image_origin.save(resultfilename, "JPEG")
#手部識別
#filename:原圖片名(本地存盤包括路徑)
def hand_analysis(filename,resultfilename,size,color,pointsize,pointcolor):
request_url = "https://aip.baidubce.com/rest/2.0/image-classify/v1/hand_analysis"
print(filename)
# 二進制方式打開圖片檔案
f = open(filename, 'rb')
img = base64.b64encode(f.read())
params = dict()
params['image'] = img
params = urllib.parse.urlencode(params).encode("utf-8")
#params = json.dumps(params).encode('utf-8')
access_token = get_token()
begin = time.perf_counter()
request_url = request_url + "?access_token=" + access_token
request = urllib.request.Request(url=request_url, data=https://www.cnblogs.com/AIBOOM/p/params)
request.add_header('Content-Type', 'application/x-www-form-urlencoded')
response = urllib.request.urlopen(request)
content = response.read()
end = time.perf_counter()
print('處理時長:'+'%.2f'%(end-begin)+'秒')
if content:
#print(content)
content=content.decode('utf-8')
#print(content)
data = https://www.cnblogs.com/AIBOOM/p/json.loads(content)
print('hand_num:',data['hand_num'])
#print(data)
result=data['hand_info']
draw_hands_point(filename,result,resultfilename,pointsize,pointcolor)
4.功能評測:
選用不同的資料對效果進行測驗,具體效果如下(以下例子均來自網上):
處理時長:0.44秒
hand_num: 1
處理時長:0.67秒
hand_num: 1
處理時長:0.56秒
hand_num: 1
處理時長:0.86秒
hand_num: 1
可以發現對于單手的情況,速度很快,效果很準確,
處理時長:0.61秒
hand_num: 3
5.測驗結論和建議
測驗下來,整體識別效果不錯,對于手部關鍵點有較強的識別能力,效果很好,速度也很快,
不過對于比較復雜的圖片,如多個手或者背景比較復雜的情況,識別率還有提高的空間,希望后續進一步提高,
作者:才能我浪費99
轉載請註明出處,本文鏈接:https://www.uj5u.com/gongcheng/5467.html
標籤:其他
