Multiscale Crowd Counting and Localization By Multitask Point Supervision

Zand, Mohsen; Haleh, Damirchi,; Farley, A. N.; Mahdiyar, Molahasani,; Greenspan, Michael; Etemad, Ali

doi:10.1109/icassp43922.2022.9747776

Cited by 32 publications

(15 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The nearest neighbor distance between prediction point set and ground-truth point set is calculated as the evaluation condition. Similar to Ribera et al ( 2019 ) and Zand et al ( 2022 ), the point is considered to belong to TP when the predicted point is within 10 pixels of some ground-truth point, otherwise it is classified to FP. Each ground-truth point is matched against only one predicted point.…”

Section: Resultsmentioning

confidence: 99%

“…To demonstrate the effectiveness of the proposed method for cotton boll localization, we compare it with several object localization networks and object detection networks on the CBL dataset. Specifically, comparisons are made with the bounding box annotation-based SSD (Liu et al, 2016 ), FasterRCNN (Ren et al, 2017 ), YOLOv3 series (Redmon and Farhadi, 2018 ), and YOLOv5 series and point annotation based object localization methods P2PNet (Song et al, 2021 ) and MSPSNet (Zand et al, 2022 ). The specific experimental results are shown in Table 1 .…”

Section: Resultsmentioning

confidence: 99%

“…Specifically, comparisons are made with the bounding box annotation-based SSD (Liu et al, 2016), FasterRCNN (Ren The optimal values in each column are bold-faced. (Redmon and Farhadi, 2018), and YOLOv5 series and point annotation based object localization methods P2PNet (Song et al, 2021) and MSPSNet (Zand et al, 2022). The specific experimental results are shown in Table 1.…”

Section: Model Evaluation Comparison Of DI Erent Localization Methodsmentioning

confidence: 99%

“…Song et al ( 2021 ) proposed the P2PNet network to directly predict a set of points and perform one-to-one matching for dense object localization. Zand et al ( 2022 ) proposed a multi-task dense object localization method based on VGG network. Although these methods can localize dense objects only with point annotations, they are difficult to localize various crops effectively due to their limited feature extraction ability, field crop scale changes, and natural plant growth changes.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Cotton boll localization method based on point annotation and multi-scale fusion

Sun

Yang

et al. 2022

Front. Plant Sci.

View full text Add to dashboard Cite

Cotton is an important source of fiber. The precise and intelligent management of cotton fields is the top priority of cotton production. Many intelligent management methods of cotton fields are inseparable from cotton boll localization, such as automated cotton picking, sustainable boll pest control, boll maturity analysis, and yield estimation. At present, object detection methods are widely used for crop localization. However, object detection methods require relatively expensive bounding box annotations for supervised learning, and some non-object regions are inevitably included in the annotated bounding boxes. The features of these non-object regions may cause misjudgment by the network model. Unlike bounding box annotations, point annotations are less expensive to label and the annotated points are only likely to belong to the object. Considering these advantages of point annotation, a point annotation-based multi-scale cotton boll localization method is proposed, called MCBLNet. It is mainly composed of scene encoding for feature extraction, location decoding for localization prediction and localization map fusion for multi-scale information association. To evaluate the robustness and accuracy of MCBLNet, we conduct experiments on our constructed cotton boll localization (CBL) dataset (300 in-field cotton boll images). Experimental results demonstrate that MCBLNet method improves by 49.4% average precision on CBL dataset compared with typically point-based localization state-of-the-arts. Additionally, MCBLNet method outperforms or at least comparable with common object detection methods.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Model Evaluation Comparison Of DI Erent Localization Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Cotton boll localization method based on point annotation and multi-scale fusion

Sun

Yang

et al. 2022

Front. Plant Sci.

View full text Add to dashboard Cite

show abstract

“…Recently point-level annotation has drawn interest in a broad range of computer vision tasks. Beside the works concerning the detection and segmentation tasks [7,29,38,2], some works adopt point-level labels to train crowd counting [50,33] models. SPTS [37] proposes to use points for the text spotting problem.…”

Section: Point-level Labels In Visual Tasksmentioning

confidence: 99%

Pointly-Supervised Panoptic Segmentation

Fan

Zhang

Tan

2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In this paper, we propose a new approach to applying pointlevel annotations for weakly-supervised panoptic segmentation. Instead of the dense pixel-level labels used by fully supervised methods, pointlevel labels only provide a single point for each target as supervision, significantly reducing the annotation burden. We formulate the problem in an end-to-end framework by simultaneously generating panoptic pseudomasks from point-level labels and learning from them. To tackle the core challenge, i.e., panoptic pseudo-mask generation, we propose a principled approach to parsing pixels by minimizing pixel-to-point traversing costs, which model semantic similarity, low-level texture cues, and high-level manifold knowledge to discriminate panoptic targets. We conduct experiments on the Pascal VOC and the MS COCO datasets to demonstrate the approach's effectiveness and show state-of-the-art performance in the weakly-supervised panoptic segmentation problem. Codes are available at https://github.com/BraveGroup/PSPS.git.

show abstract

Density Map Augmentation-Based Point-to-Point Vehicle Counting and Localization in Remote Sensing Imagery with Limited Resolution

Tan,

Guo,

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Multiscale Crowd Counting and Localization By Multitask Point Supervision

Cited by 32 publications

References 19 publications

Cotton boll localization method based on point annotation and multi-scale fusion

Cotton boll localization method based on point annotation and multi-scale fusion

Pointly-Supervised Panoptic Segmentation

Density Map Augmentation-Based Point-to-Point Vehicle Counting and Localization in Remote Sensing Imagery with Limited Resolution

Contact Info

Product

Resources

About