No detections of small objects by the tracker with DeepStream

Hi Everyone,

I have been struggling to get consistent detections on videos of single coins, while using the following:
- Tracker: nvdcf (running on deepstream 5.1)
- Model: yolov4, cspdarknet 19
- Video speed: 30fps
- Video resolution: 1920 x 1080
- Size of coins: 67 x 67

When running the model only on the video, there are detections on every frame, but with the tracker, there would only be detections when the probation age is set to 0, in which case the tracker would generate a new object id for the same object every single frame.

Are there are specific tracker parameters that would help improve the tracking? I have noticed that this only happens to small objects that move more than its own size every frame, is it a known problem?

• Hardware Platform (Jetson / GPU) Jetson Xavier NX
• DeepStream Version 5.1
• JetPack Version (valid for Jetson only) 4.5.1
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs) question
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

%YAML 1.2

  # [General]
  useUniqueID: 0 # Use 64-bit long Unique ID when assignining tracker ID. Default is [true]
  maxTargetsPerStream: 99 # 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

  # [Feature Extraction]
  useColorNames: 1 # Use ColorNames feature
  useHog: 0 # Use Histogram-of-Oriented-Gradient (HOG) feature
  useHighPrecisionFeature: 1 # Use high-precision in feature extraction. Default is [true]

  # [DCF]
  filterLr: 0.33999999999999997 # learning rate for DCF filter in exponential moving average. Valid Range: [0.0, 1.0]
  filterChannelWeightsLr: 0.31 # learning rate for the channel weights among feature channels. Valid Range: [0.0, 1.0]
  gaussianSigma: 0.19 # Standard deviation for Gaussian for desired response when creating DCF filter [pixels]
  featureImgSizeLevel: 5 # Size of a feature image. Valid range: {1, 2, 3, 4, 5}, from the smallest to the largest
  SearchRegionPaddingScale: 3 # Search region size. Determines how large the search region should be scaled from the target bbox.  Valid range: {1, 2, 3}, from the smallest to the largest

  # [MOT] [False Alarm Handling]
  maxShadowTrackingAge: 11  # Max length of shadow tracking (the shadow tracking age is incremented when (1) there's detector input yet no match or (2) tracker confidence is lower than minTrackerConfidence). Once reached, the tracker will be terminated.
  probationAge: 1       # Once the tracker age (incremented at every frame) reaches this, the tracker is considered to be valid
  earlyTerminationAge: 0  # Early termination age (in terms of shadow tracking age) during the probation period. If reached during the probation period, the tracker will be terminated prematurely.

  # [Tracker Creation Policy] [Target Candidacy]
  minDetectorConfidence: 0.3  # If the confidence of a detector bbox is lower than this, then it won't be considered for tracking
  minTrackerConfidence: 0.47  # 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]
  minTargetBboxSize: 10    # If the width or height of the bbox size gets smaller than this threshold, the target will be terminated.
  minDetectorBboxVisibilityTobeTracked: 0.8400000000000001  # If the detector-provided bbox's visibility (i.e., IOU with image) is lower than this, it won't be considered.
  minVisibiilty4Tracking: 0.9299999999999999  # If the visibility of the tracked object (i.e., IOU with image) is lower than this, it will be terminated immediately, assuming it is going out of scene.

  # [Tracker Termination Policy]
  targetDuplicateRunInterval: 3 # The interval in which the duplicate target detection removal is carried out. A Negative value indicates indefinite interval. Unit: [frames]
  minIou4TargetDuplicate: 0.72 # If the IOU of two target bboxes are higher than this, the newer target tracker will be terminated.

  # [Data Association] Matching method
  useGlobalMatching: 1   # If true, enable a global matching algorithm (i.e., Hungarian method). Otherwise, a greedy algorithm wll be used.
  usePersistentThreads: 0  # If true, create data association threads once and re-use them

  # [Data Association] Thresholds in matching scores to be considered as a valid candidate for matching
  minMatchingScore4Overall: 0.5   # Min total score
  minMatchingScore4SizeSimilarity: 0.72  # Min bbox size similarity score
  minMatchingScore4Iou: 0.55     # Min IOU score
  minMatchingScore4VisualSimilarity: 0.85  # Min visual similarity score

  # [Data Association] Weights for each matching score term
  matchingScoreWeight4VisualSimilarity: 0.9 # Weight for the visual similarity (in terms of correlation response ratio)
  matchingScoreWeight4SizeSimilarity: 0.30000000000000004   # Weight for the Size-similarity score
  matchingScoreWeight4Iou: 0.35         # Weight for the IOU score
  matchingScoreWeight4Age: 0.8         # Weight for the tracker age

  # [State Estimator]
  useTrackSmoothing: 1  # Use a state estimator
  stateEstimatorType: 2   # The type of state estimator among { moving_avg:1, kalman_filter:2 }

  # [State Estimator] [MovingAvgEstimator]
  trackExponentialSmoothingLr_loc: 0.5     # Learning rate for new location
  trackExponentialSmoothingLr_scale: 0.5  # Learning rate for new scale
  trackExponentialSmoothingLr_velocity: 0.5  # Learning rate for new velocity

  # [State Estimator] [Kalman Filter]
  kfProcessNoiseVar4Loc: 0.23  # Process noise variance for location in Kalman filter
  kfProcessNoiseVar4Scale: 0.22  # Process noise variance for scale in Kalman filter
  kfProcessNoiseVar4Vel: 0.88   # Process noise variance for velocity in Kalman filter
  kfMeasurementNoiseVar4Trk: 19.720000000000002 # Measurement noise variance for tracker's detection in Kalman filter
  kfMeasurementNoiseVar4Det: 8.72 # Measurement noise variance for detector's detection in Kalman filter

  # [Past-frame Data]
  useBufferedOutput: 1   # Enable storing of past-frame data in a buffer and report it back

  # [Instance-awareness]
  useInstanceAwareness: 1 # Use instance-awareness for multi-object tracking
  lambda_ia: 0.63      # Regularlization factor for each instance
  maxInstanceNum_ia: 4  # The number of nearby object instances to use for instance-awareness

• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

Can you have a try with NvDCF tracker with latest DeepStream version 6.2? Can you share videos to show the issue?

Hi, we have a run on deepstream 6.2, would it be possible to share the video privately? Thanks

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Yes, please share the video with private channel.

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