Loss of detected objects on enabling tracker in Deepstream-6.0 (YOLO)d

I’ve been encountering a perplexing issue while working with DeepStream 6.0 and YOLO for object detection and tracking. Specifically, I’ve noticed a significant loss of detected objects when enabling tracking within the framework. Despite experimenting with various parameters in the “config_tracker_NvDCF_max_perf.yml” file, the problem persists.

Could anyone shed some suggestions on how to mitigate this issue would be greatly appreciated. Thank you in advance for your assistance!

Here’s the config_tracker_NvDCF_max_perf.yml for reference :

minDetectorConfidence: 0.0 # If the confidence of a detector bbox is lower than this, then it won’t be considered for tracking

enableBboxUnClipping: 0 # In case the bbox is likely to be clipped by image border, unclip bbox
maxTargetsPerStream: 100 # Max number of targets to track per stream. Recommended to set >10. Note: this value should account for the targets being tracked in shadow mode as well. Max value depends on the GPU memory capacity

[Creation & Termination Policy]

minIouDiff4NewTarget: 0.1 # If the IOU between the newly detected object and any of the existing targets is higher than this threshold, this newly detected object will be discarded.
minTrackerConfidence: 0.1 # If the confidence of an object tracker is lower than this on the fly, then it will be tracked in shadow mode. Valid Range: [0.0, 1.0]
probationAge: 0.1 # If the target’s age exceeds this, the target will be considered to be valid.
maxShadowTrackingAge: 100 # Max length of shadow tracking. If the shadowTrackingAge exceeds this limit, the tracker will be terminated.
earlyTerminationAge: 10 # If the shadowTrackingAge reaches this threshold while in TENTATIVE period, the target will be terminated prematurely.

useUniqueID: 0 # Use 64-bit long Unique ID when assignining tracker ID.

dataAssociatorType: 0 # the type of data associator among { DEFAULT= 0 }
associationMatcherType: 0 # the type of matching algorithm among { GREEDY=0, GLOBAL=1 }
checkClassMatch: 0 # If checked, only the same-class objects are associated with each other. Default: true

[Association Metric: Thresholds for valid candidates]

minMatchingScore4Overall: 0.0 # Min total score
minMatchingScore4SizeSimilarity: 0.2 # Min bbox size similarity score
minMatchingScore4Iou: 0.0 # Min IOU score
minMatchingScore4VisualSimilarity: 0.2 # Min visual similarity score

[Association Metric: Weights]

matchingScoreWeight4VisualSimilarity: 0.2 # Weight for the visual similarity (in terms of correlation response ratio)
matchingScoreWeight4SizeSimilarity: 0.0 # Weight for the Size-similarity score
matchingScoreWeight4Iou: 0.2 # Weight for the IOU score

stateEstimatorType: 1 # the type of state estimator among { DUMMY=0, SIMPLE=1, REGULAR=2 }

[Dynamics Modeling]

processNoiseVar4Loc: 2.0 # Process noise variance for bbox center
processNoiseVar4Size: 1.0 # Process noise variance for bbox size
processNoiseVar4Vel: 0.1 # Process noise variance for velocity
measurementNoiseVar4Detector: 1.0 # Measurement noise variance for detector’s detection
measurementNoiseVar4Tracker: 1.0 # Measurement noise variance for tracker’s localization

visualTrackerType: 1 # the type of visual tracker among { DUMMY=0, NvDCF=1 }

[NvDCF: Feature Extraction]

useColorNames: 1 # Use ColorNames feature
useHog: 0 # Use Histogram-of-Oriented-Gradient (HOG) feature
featureImgSizeLevel: 1 # Size of a feature image. Valid range: {1, 2, 3, 4, 5}, from the smallest to the largest
featureFocusOffsetFactor_y: -0.2 # The offset for the center of hanning window relative to the feature height. The center of hanning window would move by (featureFocusOffsetFactor_y*featureMatSize.height) in vertical direction

[NvDCF: Correlation Filter]

filterLr: 0.075 # learning rate for DCF filter in exponential moving average. Valid Range: [0.0, 1.0]
filterChannelWeightsLr: 0.1 # learning rate for the channel weights among feature channels. Valid Range: [0.0, 1.0]
gaussianSigma: 0.75 # Standard deviation for Gaussian for desired response when creating DCF filter [pixels]

I see so many object in your picture. Can you ensure maxTargetsPerStream is enough for your use case?

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.