A Normalized Gaussian Wasserstein Distance for Tiny Object Detection

• URL: http://arxiv.org/abs/2110.13389v1
• Date: Tue, 26 Oct 2021 03:43:17 GMT
• Title: A Normalized Gaussian Wasserstein Distance for Tiny Object Detection
• Authors: Jinwang Wang, Chang Xu, Wen Yang, Lei Yu
• Abstract summary: State-of-the-art detectors do not produce satisfactory results on tiny objects due to lack of appearance information. We propose a new evaluation metric using Wasserstein distance for tiny object detection.
• Score: 21.50800336432307
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Detecting tiny objects is a very challenging problem since a tiny object only contains a few pixels in size. We demonstrate that state-of-the-art detectors do not produce satisfactory results on tiny objects due to the lack of appearance information. Our key observation is that Intersection over Union (IoU) based metrics such as IoU itself and its extensions are very sensitive to the location deviation of the tiny objects, and drastically deteriorate the detection performance when used in anchor-based detectors. To alleviate this, we propose a new evaluation metric using Wasserstein distance for tiny object detection. Specifically, we first model the bounding boxes as 2D Gaussian distributions and then propose a new metric dubbed Normalized Wasserstein Distance (NWD) to compute the similarity between them by their corresponding Gaussian distributions. The proposed NWD metric can be easily embedded into the assignment, non-maximum suppression, and loss function of any anchor-based detector to replace the commonly used IoU metric. We evaluate our metric on a new dataset for tiny object detection (AI-TOD) in which the average object size is much smaller than existing object detection datasets. Extensive experiments show that, when equipped with NWD metric, our approach yields performance that is 6.7 AP points higher than a standard fine-tuning baseline, and 6.0 AP points higher than state-of-the-art competitors.

Related papers

• Similarity Distance-Based Label Assignment for Tiny Object Detection [17.059514012235354]
  We introduce a simple but effective strategy named the Similarity Distance (SimD) to evaluate the similarity between bounding boxes. Our approach can be applied in common anchor-based detectors in place of the IoU for label assignment and Non Maximum Suppression (NMS) Experiments on four mainstream tiny object detection datasets demonstrate superior performance of our method.
  arXiv  Detail & Related papers  (2024-07-02T16:12:37Z)
• Visible and Clear: Finding Tiny Objects in Difference Map [50.54061010335082]
  We introduce a self-reconstruction mechanism in the detection model, and discover the strong correlation between it and the tiny objects. Specifically, we impose a reconstruction head in-between the neck of a detector, constructing a difference map of the reconstructed image and the input, which shows high sensitivity to tiny objects. We further develop a Difference Map Guided Feature Enhancement (DGFE) module to make the tiny feature representation more clear.
  arXiv  Detail & Related papers  (2024-05-18T12:22:26Z)
• Size-invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection [133.66006666465447]
  Current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. We develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes.
  arXiv  Detail & Related papers  (2024-05-16T03:01:06Z)
• YOLC: You Only Look Clusters for Tiny Object Detection in Aerial Images [33.80392696735718]
  YOLC (You Only Look Clusters) is an efficient and effective framework that builds on an anchor-free object detector, CenterNet. To overcome the challenges posed by large-scale images and non-uniform object distribution, we introduce a Local Scale Module (LSM) that adaptively searches cluster regions for zooming in for accurate detection. We perform extensive experiments on two aerial image datasets, including Visdrone 2019 and UAVDT, to demonstrate the effectiveness and superiority of our proposed approach.
  arXiv  Detail & Related papers  (2024-04-09T10:03:44Z)
• Small Object Detection via Coarse-to-fine Proposal Generation and Imitation Learning [52.06176253457522]
  We propose a two-stage framework tailored for small object detection based on the Coarse-to-fine pipeline and Feature Imitation learning. CFINet achieves state-of-the-art performance on the large-scale small object detection benchmarks, SODA-D and SODA-A.
  arXiv  Detail & Related papers  (2023-08-18T13:13:09Z)
• RFLA: Gaussian Receptive Field based Label Assignment for Tiny Object Detection [45.10513110142015]
  Current anchor-based or anchor-free label assignment paradigms incur many outlier tiny-sized ground truth samples. We propose a Gaussian Receptive Field based Label Assignment (RFLA) strategy for tiny object detection. Our approach outperforms the state-of-the-art competitors with 4.0 AP points on the AI-TOD dataset.
  arXiv  Detail & Related papers  (2022-08-18T09:35:56Z)
• Detecting tiny objects in aerial images: A normalized Wasserstein distance and a new benchmark [45.10513110142015]
  We propose a new evaluation metric dubbed Normalized Wasserstein Distance (NWD) and a new RanKing-based Assigning (RKA) strategy for tiny object detection. The proposed NWD-RKA strategy can be easily embedded into all kinds of anchor-based detectors to replace the standard IoU threshold-based one. Tested on four datasets, NWD-RKA can consistently improve tiny object detection performance by a large margin.
  arXiv  Detail & Related papers  (2022-06-28T13:33:06Z)
• EPP-Net: Extreme-Point-Prediction-Based Object Detection [9.270523894683278]
  We present a new anchor-free dense object detector, which regresses the relative displacement vector between each pixel and the four extreme points. We also propose a new metric to measure the similarity between two groups of extreme points, namely, Extreme Intersection over Union (EIoU) On the MS-COCO dataset, our method achieves an average precision (AP) of 39.3% with ResNet-50 and an AP of 48.3% with ResNeXt-101-DCN.
  arXiv  Detail & Related papers  (2021-04-29T01:01:50Z)
• SCRDet++: Detecting Small, Cluttered and Rotated Objects via Instance-Level Feature Denoising and Rotation Loss Smoothing [131.04304632759033]
  Small and cluttered objects are common in real-world which are challenging for detection. In this paper, we first innovatively introduce the idea of denoising to object detection. Instance-level denoising on the feature map is performed to enhance the detection to small and cluttered objects.
  arXiv  Detail & Related papers  (2020-04-28T06:03:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.