Deepstream-test2 returns partial frames list of track

amir.friedman · January 23, 2024, 9:11am

Hardware Platform : GPU
DeepStream Version 6.3

TensorRT Version 8.6.1.6
NVIDIA GPU Driver Version 535.146.02

When running deepstream-test2 (python) with the default config file and no modifications, I noticed that I am only getting partial results. Specifically, the frames list of each track contains only the first three frames.
Can you provide insights or suggestions on why this might be happening?

junshengy · January 24, 2024, 10:43am

How did you run it? Can you provide the command line and detailed error log?

Also which GPU are you using and whether you have a monitor ？

amir.friedman · January 28, 2024, 11:02am

Hello,

My GPU: NVIDIA GeForce RTX 3070

I ran:
python3 deepstream_test_2.py <my_clip>

Upon observation, I notice that there is no error, and everything appears to be functioning correctly.
However, when I observe four pedestrians walking for approximately one minute, I observe that only the first three bounding boxes of each track are present in the output.

I would appreciate your guidance on how to obtain the complete set of bounding boxes for each tracked pedestrian.
If there are any settings or configurations that need adjustment, please provide instructions on how to address this issue.

Thank you for your prompt attention to this matter.

Best regards,
Amir

amir.friedman · January 28, 2024, 11:08am

The output:
INFO: …/nvdsinfer/nvdsinfer_model_builder.cpp:610 [Implicit Engine Info]: layers num: 2
*0 INPUT kFLOAT input_1 3x224x224 *
*1 OUTPUT kFLOAT predictions/Softmax 6x1x1 *

INFO: …/nvdsinfer/nvdsinfer_model_builder.cpp:610 [Implicit Engine Info]: layers num: 2
*0 INPUT kFLOAT input_1 3x224x224 *
*1 OUTPUT kFLOAT predictions/Softmax 20x1x1 *

INFO: …/nvdsinfer/nvdsinfer_model_builder.cpp:610 [Implicit Engine Info]: layers num: 2
*0 INPUT kFLOAT input_1 3x224x224 *
*1 OUTPUT kFLOAT predictions/Softmax 12x1x1 *

gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
[NvMultiObjectTracker] Initialized
*Creating Pipeline *

*Creating Source *

*Creating H264Parser *

*Creating Decoder *

*Creating EGLSink *

*Playing file /mnt/video-db/Test_Videos/NotConstantFPS/Short/ABIA-Escalator1_S1.avi *
*Adding elements to Pipeline *

*Linking elements in the Pipeline *

*Starting pipeline *

Frame Number=0 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=1 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=2 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=3 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
numobj= 3
uniqueId= 0
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1653.0693359375
tBbox.width: 120.945556640625
tBbox.top: 472.7581787109375
tBbox.right: 286.55621337890625
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1654.6241455078125
tBbox.width: 120.38873291015625
tBbox.top: 466.0554504394531
tBbox.right: 290.51348876953125
confidence: 0.903741180896759
age: 2
frameNum: 2
tBbox.left: 1658.116943359375
tBbox.width: 121.8167724609375
tBbox.top: 460.907958984375
tBbox.right: 284.2696228027344
confidence: 0.7915024161338806
age: 3
numobj= 3
uniqueId= 1
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1240.6007080078125
tBbox.width: 143.51998901367188
tBbox.top: 607.5460205078125
tBbox.right: 308.93621826171875
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1247.601318359375
tBbox.width: 135.25204467773438
tBbox.top: 632.09423828125
tBbox.right: 280.7585754394531
confidence: 0.9064123034477234
age: 2
frameNum: 2
tBbox.left: 1250.189453125
tBbox.width: 133.256103515625
tBbox.top: 631.6926879882812
tBbox.right: 280.4739074707031
confidence: 0.7988146543502808
age: 3
numobj= 3
uniqueId= 2
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1584.93310546875
tBbox.width: 50.120361328125
tBbox.top: 94.63676452636719
tBbox.right: 113.76005554199219
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1585.00439453125
tBbox.width: 49.99090576171875
tBbox.top: 94.74036407470703
tBbox.right: 113.58354949951172
confidence: 0.946975588798523
age: 2
frameNum: 2
tBbox.left: 1590.3056640625
tBbox.width: 46.89532470703125
tBbox.top: 94.89263916015625
tBbox.right: 110.53397369384766
confidence: 0.9286273717880249
age: 3
numobj= 3
uniqueId= 3
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1524.3753662109375
tBbox.width: 44.5693359375
tBbox.top: 95.95864868164062
tBbox.right: 107.1131820678711
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1524.490234375
tBbox.width: 44.646697998046875
tBbox.top: 95.87523651123047
tBbox.right: 107.1480712890625
confidence: 0.9225800633430481
age: 2
frameNum: 2
tBbox.left: 1524.4886474609375
tBbox.width: 44.370758056640625
tBbox.top: 96.14335632324219
tBbox.right: 106.9726333618164
confidence: 0.8891088366508484
age: 3
Frame Number=4 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=5 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=6 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=7 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=8 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=9 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=10 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=11 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=12 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=13 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=14 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=15 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=16 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=17 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=18 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=19 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=20 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=21 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=22 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=23 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=24 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=25 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=26 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=27 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=28 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=29 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=30 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=31 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=32 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=33 Number of Objects=4 Vehicle_count=0 Person_count=4

Best regards,
Amir

junshengy · January 29, 2024, 6:24am

Is this your test stream? Can you share it? I’m not sure if you need to change the configuration file.

By the way, we use sample_720p.h264 as the test stream, only 48s

amir.friedman · January 31, 2024, 10:01am

nv.zip (33.8 MB)

amir.friedman · January 31, 2024, 10:02am

Hi,

Attched zip with the clip

Best regards,
Amir

locybdxlyb204 · February 2, 2024, 8:26am

Can you share your code where you get the information of objects?

junshengy · February 6, 2024, 11:22am

We will look into this issue and will be back once there is any progress.

amir.friedman · February 18, 2024, 10:08am

I ran deepstream_test_2.py with the default config file and no modifications.
The output is printed to the screen automatically like the output attached above.

Best regards,
Amir

junshengy · March 5, 2024, 12:48pm

Do you mean that at about 1 minute, the two people on the right are missing bboxes?
Here are my results, not really sure what your problem is

amir.friedman · March 6, 2024, 10:28am

I’m experiencing an issue with the printed output where only the first three frames of each object are being displayed, despite the fact that in the movie, each bbox appears in many more than 3 frames.
The output:

Frame Number=0 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=1 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=2 Number of Objects=0 Vehicle_count=0 Person_count=0
streamId= 0
surfaceStreamID= 0
Frame Number=3 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
numobj= 3
uniqueId= 0
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1653.0693359375
tBbox.width: 120.945556640625
tBbox.top: 472.7581787109375
tBbox.right: 286.55621337890625
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1654.6241455078125
tBbox.width: 120.38873291015625
tBbox.top: 466.0554504394531
tBbox.right: 290.51348876953125
confidence: 0.903741180896759
age: 2
frameNum: 2
tBbox.left: 1658.116943359375
tBbox.width: 121.8167724609375
tBbox.top: 460.907958984375
tBbox.right: 284.2696228027344
confidence: 0.7915024161338806
age: 3
numobj= 3
uniqueId= 1
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1240.6007080078125
tBbox.width: 143.51998901367188
tBbox.top: 607.5460205078125
tBbox.right: 308.93621826171875
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1247.601318359375
tBbox.width: 135.25204467773438
tBbox.top: 632.09423828125
tBbox.right: 280.7585754394531
confidence: 0.9064123034477234
age: 2
frameNum: 2
tBbox.left: 1250.189453125
tBbox.width: 133.256103515625
tBbox.top: 631.6926879882812
tBbox.right: 280.4739074707031
confidence: 0.7988146543502808
age: 3
numobj= 3
uniqueId= 2
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1584.93310546875
tBbox.width: 50.120361328125
tBbox.top: 94.63676452636719
tBbox.right: 113.76005554199219
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1585.00439453125
tBbox.width: 49.99090576171875
tBbox.top: 94.74036407470703
tBbox.right: 113.58354949951172
confidence: 0.946975588798523
age: 2
frameNum: 2
tBbox.left: 1590.3056640625
tBbox.width: 46.89532470703125
tBbox.top: 94.89263916015625
tBbox.right: 110.53397369384766
confidence: 0.9286273717880249
age: 3
numobj= 3
uniqueId= 3
classId= 2
objLabel= Person
frameNum: 0
tBbox.left: 1524.3753662109375
tBbox.width: 44.5693359375
tBbox.top: 95.95864868164062
tBbox.right: 107.1131820678711
confidence: 1.0
age: 1
frameNum: 1
tBbox.left: 1524.490234375
tBbox.width: 44.646697998046875
tBbox.top: 95.87523651123047
tBbox.right: 107.1480712890625
confidence: 0.9225800633430481
age: 2
frameNum: 2
tBbox.left: 1524.4886474609375
tBbox.width: 44.370758056640625
tBbox.top: 96.14335632324219
tBbox.right: 106.9726333618164
confidence: 0.8891088366508484
age: 3
Frame Number=4 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=5 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=6 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=7 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=8 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=9 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=10 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=11 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=12 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=13 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=14 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=15 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=16 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=17 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=18 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=19 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=20 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=21 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=22 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=23 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0
surfaceStreamID= 0
Frame Number=24 Number of Objects=4 Vehicle_count=0 Person_count=4
streamId= 0

Best regards,
Amir

Topic		Replies	Views
Only results of the first 4 frames displayed while running deepstream-yolo-app DeepStream SDK	1	669	November 26, 2018
PERF {stream:0.0} Error After 33 Frames in DeepStream Application DeepStream SDK deepstream	3	92	December 30, 2024
DeepStream test 3 sample app processes more frames than the number of actual video frames DeepStream SDK	9	547	July 22, 2022
Same pipeline, run twice, different results !!!! DeepStream SDK	5	850	October 12, 2021
Where is problem of test4 pipline? DeepStream SDK jetson-inference	36	2024	February 1, 2022
Deepstream Segmentaion fault DeepStream SDK	6	568	February 28, 2024
Deepstream objects detected boxes are continuously highly moving around the detected objects. DeepStream SDK	18	3324	October 12, 2021
deepstream : deepstream_test1 not detecting Jetson AGX Xavier	2	1228	October 18, 2021
Custom_parse_bbox_func fails to send data to deepstream for some frames DeepStream SDK tensorrt , python , inference-server-triton , deepstream , deepstream61	5	912	April 6, 2023
Adding tracker to deepstream_python_apps/apps/deepstream-test3 DeepStream SDK python , inference-server-triton , deepstream	9	1475	February 21, 2023

Deepstream-test2 returns partial frames list of track

Related topics