Styjun

Published paper containing well-known results

Does entangle require vegetation?

Why is dry soil hydrophobic? Bad gardener paradox

Confused about 誘われて (Sasowarete)

Why would an Inquisitive rogue choose to use Insightful Fighting as opposed to using their Cunning Action to Hide?

Mistakenly modified `/bin/sh'

What is the English equivalent of 干物女 (dried fish woman)?

Why are Japanese translated subtitles non-conversational?

Why would guns not work in the dungeon?

Evaluating an integral symbolically

Meaning of slash chord without anything left of the slash

Bob's unnecessary trip to the shops

Why does the autopilot disengage even when it does not receive pilot input?

Too many spies!

Are there any double stars that I can actually see orbit each other?

How can an advanced civilization forget how to manufacture its technology?

Why does Hellboy file down his horns?

What is this old "lemon-squeezer" shaped pan

Why do they not say "The Baby"

What's the difference between soft PWM and PWM

Do native speakers use ZVE or CPU?

Integral clarification

Add region constraint to Graphics

Why does the trade federation become so alarmed upon learning the ambassadors are Jedi Knights?

pool map no performance gain

Python reverse / invert a mappingList comprehension vs mapKeyboard Interrupts with python's multiprocessing PoolThreading pool similar to the multiprocessing Pool?Show the progress of a Python multiprocessing pool map call?multiprocessing.Pool: When to use apply, apply_async or map?Python multiprocessing: no performance gain with multiple processesHow to read in a video file with cv2.VideoCapture?Segmentation fault OpenCV cap.read udp stream PythonMove a function into a thread for faster process in python

.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty margin-bottom:0;

2

I'm doing video capture from opencv using the following

frame_count = 90000
 while f < frame_count:
 ret, frame = cap.read()
 f+=1
 if (f > 41300):
 dst = cv2.warpPerspective(frame.copy(), matrix, (frame.shape[1], frame.shape[0]))
 pipeline(pool_1, pool_2, frame, car_tracker, ped_tracker, df_region, region_buffered, df_line_car, df_line_ped, det, dst, matrix)
 cv2.imshow("frame", frame)
 cv2.imshow('dst', dst)
 cv2.waitKey(1)

In function pipeline:

df_cars= pool_1.map(compute, cars)
df_peds = pool_2.map(compute, peds)

compute is:

def compute(v):
 gb = ('trajectory_id',)
 global general_pd
 id = v[0]
 x = v[1]
 y = v[2]
 frame_id = v[3]
 general_pd.loc[len(general_pd)] = [id, x,y, frame_id]
 grouped = general_pd.loc[general_pd['trajectory_id'] == id]
 df_region = v[4]
 region_buffered = v[5]
 df_line = v[6]
 rpp = RawParameterProcessor(grouped, df_line, frame_idx, df_region, region_buffered, gb=gb)
 df_parameter_car = rpp.compute()
 return df_parameter_car

I don't know what I'm doing wrong, so that I can't get the correct performance gain, I launch two process, and each frame of the video capture I launch a process and do the jobs async, but I don't get any performance gain.

python multiprocessing

share|improve this question

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  can you show the type of the pool objects?
  
  – Jean-François Fabre♦
  Mar 26 at 6:14
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  pool_1 = Pool(2) pool_2 = Pool(2)
  
  – waleed saleh
  Mar 26 at 6:16
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  from multiprocessing import Pool, Queue
  
  – waleed saleh
  Mar 26 at 6:17
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @Jean-FrançoisFabre det is tensor flow detector, is that the problem ?
  
  – waleed saleh
  Mar 26 at 6:29
  

  
  
  
  

add a comment | 

2

I'm doing video capture from opencv using the following

frame_count = 90000
 while f < frame_count:
 ret, frame = cap.read()
 f+=1
 if (f > 41300):
 dst = cv2.warpPerspective(frame.copy(), matrix, (frame.shape[1], frame.shape[0]))
 pipeline(pool_1, pool_2, frame, car_tracker, ped_tracker, df_region, region_buffered, df_line_car, df_line_ped, det, dst, matrix)
 cv2.imshow("frame", frame)
 cv2.imshow('dst', dst)
 cv2.waitKey(1)

In function pipeline:

df_cars= pool_1.map(compute, cars)
df_peds = pool_2.map(compute, peds)

compute is:

def compute(v):
 gb = ('trajectory_id',)
 global general_pd
 id = v[0]
 x = v[1]
 y = v[2]
 frame_id = v[3]
 general_pd.loc[len(general_pd)] = [id, x,y, frame_id]
 grouped = general_pd.loc[general_pd['trajectory_id'] == id]
 df_region = v[4]
 region_buffered = v[5]
 df_line = v[6]
 rpp = RawParameterProcessor(grouped, df_line, frame_idx, df_region, region_buffered, gb=gb)
 df_parameter_car = rpp.compute()
 return df_parameter_car

I don't know what I'm doing wrong, so that I can't get the correct performance gain, I launch two process, and each frame of the video capture I launch a process and do the jobs async, but I don't get any performance gain.

python multiprocessing

share|improve this question

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  can you show the type of the pool objects?
  
  – Jean-François Fabre♦
  Mar 26 at 6:14
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  pool_1 = Pool(2) pool_2 = Pool(2)
  
  – waleed saleh
  Mar 26 at 6:16
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  from multiprocessing import Pool, Queue
  
  – waleed saleh
  Mar 26 at 6:17
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @Jean-FrançoisFabre det is tensor flow detector, is that the problem ?
  
  – waleed saleh
  Mar 26 at 6:29
  

  
  
  
  

add a comment | 

I'm doing video capture from opencv using the following

frame_count = 90000
 while f < frame_count:
 ret, frame = cap.read()
 f+=1
 if (f > 41300):
 dst = cv2.warpPerspective(frame.copy(), matrix, (frame.shape[1], frame.shape[0]))
 pipeline(pool_1, pool_2, frame, car_tracker, ped_tracker, df_region, region_buffered, df_line_car, df_line_ped, det, dst, matrix)
 cv2.imshow("frame", frame)
 cv2.imshow('dst', dst)
 cv2.waitKey(1)

In function pipeline:

df_cars= pool_1.map(compute, cars)
df_peds = pool_2.map(compute, peds)

compute is:

def compute(v):
 gb = ('trajectory_id',)
 global general_pd
 id = v[0]
 x = v[1]
 y = v[2]
 frame_id = v[3]
 general_pd.loc[len(general_pd)] = [id, x,y, frame_id]
 grouped = general_pd.loc[general_pd['trajectory_id'] == id]
 df_region = v[4]
 region_buffered = v[5]
 df_line = v[6]
 rpp = RawParameterProcessor(grouped, df_line, frame_idx, df_region, region_buffered, gb=gb)
 df_parameter_car = rpp.compute()
 return df_parameter_car

I don't know what I'm doing wrong, so that I can't get the correct performance gain, I launch two process, and each frame of the video capture I launch a process and do the jobs async, but I don't get any performance gain.

python multiprocessing

share|improve this question

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

frame_count = 90000
 while f < frame_count:
 ret, frame = cap.read()
 f+=1
 if (f > 41300):
 dst = cv2.warpPerspective(frame.copy(), matrix, (frame.shape[1], frame.shape[0]))
 pipeline(pool_1, pool_2, frame, car_tracker, ped_tracker, df_region, region_buffered, df_line_car, df_line_ped, det, dst, matrix)
 cv2.imshow("frame", frame)
 cv2.imshow('dst', dst)
 cv2.waitKey(1)

df_cars= pool_1.map(compute, cars)
df_peds = pool_2.map(compute, peds)

def compute(v):
 gb = ('trajectory_id',)
 global general_pd
 id = v[0]
 x = v[1]
 y = v[2]
 frame_id = v[3]
 general_pd.loc[len(general_pd)] = [id, x,y, frame_id]
 grouped = general_pd.loc[general_pd['trajectory_id'] == id]
 df_region = v[4]
 region_buffered = v[5]
 df_line = v[6]
 rpp = RawParameterProcessor(grouped, df_line, frame_idx, df_region, region_buffered, gb=gb)
 df_parameter_car = rpp.compute()
 return df_parameter_car

share|improve this question

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

share|improve this question

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

edited Mar 26 at 6:07

Arkistarvh Kltzuonstev

3,38622 gold badges1313 silver badges3535 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges

asked Mar 26 at 6:03

waleed salehwaleed saleh

1455 bronze badges