Bottom-up Object Detection by Grouping Extreme and Center Points

Zhuo, Jiacheng

doi:10.48550/arxiv.1901.08043

Cited by 29 publications

(11 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In (Duan et al, 2019), an extension of CornerNet is proposed by adding the center of the bounding box as keypoint. In (Zhou et al, 2019b), ExtremeNet is presented where keypoints are given by objects' extreme point. Extreme points have the advantage over corners of being always part of the object, without being affecting by background information.…”

Section: State Of the Artmentioning

confidence: 99%

Unsupervised Domain Adaptation from Synthetic to Real Images for Anchorless Object Detection

Scheck,

Grassi,

Hirtz

2020

Preprint

View full text Add to dashboard Cite

Synthetic images are one of the most promising solutions to avoid high costs associated with generating annotated datasets to train supervised convolutional neural networks (CNN). However, to allow networks to generalize knowledge from synthetic to real images, domain adaptation methods are necessary. This paper implements unsupervised domain adaptation (UDA) methods on an anchorless object detector. Given their good performance, anchorless detectors are increasingly attracting attention in the field of object detection. While their results are comparable to the well-established anchor-based methods, anchorless detectors are considerably faster. In our work, we use CenterNet, one of the most recent anchorless architectures, for a domain adaptation problem involving synthetic images. Taking advantage of the architecture of anchorless detectors, we propose to adjust two UDA methods, viz., entropy minimization and maximum squares loss, originally developed for segmentation, to object detection. Our results show that the proposed UDA methods can increase the mAP from 61 % to 69 % with respect to direct transfer on the considered anchorless detector. The code is available: https://github.com/scheckmedia/centernet-uda.

show abstract

Section: State Of the Artmentioning

confidence: 99%

Unsupervised Domain Adaptation from Synthetic to Real Images for Anchorless Object Detection

Scheck,

Grassi,

Hirtz

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The grouping method of ExtremeNet [26] is to match 4 extreme points with the center point, and basic rules are based on the assumption that if geometric center points of 4 extreme points match the predicted center points then the same object should they belong to. However, this method is still unable to be applied to detect aircraft due to the fact that many of them may be spatially symmetric in remote sensing images.…”

Section: Bottom-up Detectors Based On Dcnnsmentioning

confidence: 99%

“…One-stage detectors YOLO9000 [17] DarkNet-19 544×544 82.4% 81.4% 81.9% YOLOv3 [18] DarkNet-53 608×608 85.4% 85.5% 85.4% SSD [14] ResNet-101 513×513 88.7% 89.4% 89.0% RetinaNet [15] ResNet-101 512×512 89.8% 90.9% 90.3% ConerNet(S) [16] 104-Hourglass 511×511 76.2% 77.4% 76.8% ConerNet(M) [16] 104-Hourglass 511×511 76.8% 77.4% 77.1% ExtremeNet [26] 104 4.4.4. Error analysis X-LineNet has three forms of detection results.…”

Section: Comparison On Different Train Setmentioning

confidence: 99%

X-LineNet: Detecting Aircraft in Remote Sensing Images by a pair of Intersecting Line Segments

Wei¹,

Zhang²,

Wang³

et al. 2019

Preprint

View full text Add to dashboard Cite

In the field of aircraft detection, tremendous progress has been gained movitated by the development of deep convolution neural networks(DCNNs). At present, most state-of-art models based on DCNNs belong to top-down approaches which take a wide use of anchor mechanism. The obtaining of high accuracy in them relys on the enumeration of massive potentional locations of objects with the form of rectangular bounding box, which is wasteful and less elaborate. In this paper, we present a novel aircraft detection model in a bottom-up manner, which formulated the task as detection of two intersecting line segments inside each target and grouping of them, thus we name it as X-LineNet. As the result of learning more delicate visual grammars information of aircraft, detection results with more concrete details and higher accuracy can be gained by X-LineNet. Just for these advantages, we designed a novel form of detetction results-pentagonal mask which has less redundancy and can better represent airplanes than that of rectangular box in remote sensing images.

show abstract

“…˜100k ), which may result in easy negatives dominated training. Most recently, anchor-free detectors [16,17,18,19,20,21] gain much attention due to their simplicity, but they are driven by key-point detection (e.g. the center point), thus still suffering from the similar imbalance.…”

Section: Introductionmentioning

confidence: 99%

“…Focal Loss [13] and GHM [22]. Similarly, anchor-free detectors [17,18,20,21] apply Focal Loss or its variants for key-point prediction. Despite being effective, these schemes are usually heuristic and demand laborious hyper-parameters tuning.…”

Section: Introductionmentioning

confidence: 99%

Residual Objectness for Imbalance Reduction

Chen¹,

Liu²,

Luo³

et al. 2019

Preprint

View full text Add to dashboard Cite

For a long time, object detectors have suffered from extreme imbalance between foregrounds and backgrounds. While several sampling/reweighting schemes have been explored to alleviate the imbalance, they are usually heuristic and demand laborious hyper-parameters tuning, which is hard to achieve the optimality. In this paper, we first reveal that such the imbalance could be addressed in a learning-based manner. Guided by this illuminating observation, we propose a novel Residual Objectness (ResObj) mechanism that addresses the imbalance by end-to-end optimization, while no further hand-crafted sampling/reweighting is required. Specifically, by applying multiple cascaded objectness-related modules with residual connections, we formulate an elegant consecutive refinement procedure for distinguishing the foregrounds from backgrounds, thereby progressively addressing the imbalance. Extensive experiments present the effectiveness of our method, as well as its compatibility and adaptivity for both region-based and one-stage detectors, namely, the RetinaNet-ResObj, YOLOv3-ResObj and FasterRCNN-ResObj achieve relative 3.6%, 3.9%, 3.2% Average Precision (AP) improvements compared with their vanilla models on COCO, respectively.

show abstract

Bottom-up Object Detection by Grouping Extreme and Center Points

Cited by 29 publications

References 34 publications

Unsupervised Domain Adaptation from Synthetic to Real Images for Anchorless Object Detection

Unsupervised Domain Adaptation from Synthetic to Real Images for Anchorless Object Detection

X-LineNet: Detecting Aircraft in Remote Sensing Images by a pair of Intersecting Line Segments

Residual Objectness for Imbalance Reduction

Contact Info

Product

Resources

About