BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation

Chen, Hao; Sun, Kunyang; Tian, Zhi; Shen, Chunhua; Huang, Yongming; Yan, Youliang

doi:10.1109/cvpr42600.2020.00860

Cited by 470 publications

(182 citation statements)

References 10 publications

Supporting

Mentioning

139

Contrasting

Order By: Relevance

“…The bottom level corresponds to more detailed information, such as the position and centre of the weed, which can retain better position information. BlendMask combines the concepts of the top-down and bottom-up methodologies, thereby combining rich instance-level information with accurate dense pixel features [ 26 ].…”

Section: Methodsmentioning

confidence: 99%

“…The researchers used the improved Mask R-CNN model to obtain the phenotype of apple flowers and strawberries and achieved good segmentation under occlusion and overlap conditions [ 23 , 24 ]. The BlendMask model proposed by Chen et al exhibited a higher segmentation performance on the COCO dataset [ 25 ] than the Mask R-CNN [ 26 ]. BlendMask combines the ideas of top-down and bottom-up methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Instance Segmentation-Based Method to Obtain the Leaf Age and Plant Centre of Weeds in Complex Field Environments

Quan

Mao

et al. 2021

Sensors

View full text Add to dashboard Cite

Leaf age and plant centre are important phenotypic information of weeds, and accurate identification of them plays an important role in understanding the morphological structure of weeds, guiding precise targeted spraying and reducing the use of herbicides. In this work, a weed segmentation method based on BlendMask is proposed to obtain the phenotypic information of weeds under complex field conditions. This study collected images from different angles (front, side, and top views) of three kinds of weeds (Solanum nigrum, barnyard grass (Echinochloa crus-galli), and Abutilon theophrasti Medicus) in a maize field. Two datasets (with and without data enhancement) and two backbone networks (ResNet50 and ResNet101) were replaced to improve model performance. Finally, seven evaluation indicators are used to evaluate the segmentation results of the model under different angles. The results indicated that data enhancement and ResNet101 as the backbone network could enhance the model performance. The F1 value of the plant centre is 0.9330, and the recognition accuracy of leaf age can reach 0.957. The mIOU value of the top view is 0.642. Therefore, deep learning methods can effectively identify weed leaf age and plant centre, which is of great significance for variable spraying.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

An Instance Segmentation-Based Method to Obtain the Leaf Age and Plant Centre of Weeds in Complex Field Environments

Quan

Mao

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The most promising method seems to be the light versions of modern one-step fully convolutional approaches to the segmentation of objects Blendmask [25] and SOLOv2 [28]. They will also be explored in this article.…”

Section: Methodology a Real-time Instance Segmentation Of Indoor Scenesmentioning

confidence: 99%

“…Several modern models do not use intermediate bounding box detection but perform segmentation directly, albeit in several stages. These methods include modern architectures BlendMask [25] and CenterMask [26].…”

Section: B Real-time Object Segmentationmentioning

confidence: 99%

Real-Time Object Navigation With Deep Neural Networks and Hierarchical Reinforcement Learning

et al. 2020

View full text Add to dashboard Cite

In the last years, deep learning and reinforcement learning methods have significantly improved mobile robots in such fields as perception, navigation, and planning. But there are still gaps in applying these methods to real robots due to the low computational efficiency of recent neural network architectures and their poor adaptability to robotic experiments' realities. In this paper, we consider an important task in mobile robotics -navigation to an object using an RGB-D camera. We develop a new neural network framework for robot control that is fast and resistant to possible noise in sensors and actuators. We propose an original integration of semantic segmentation, mapping, localization, and reinforcement learning methods to improve the effectiveness of exploring the environment, finding the desired object, and quickly navigating to it. We created a new HISNav dataset based on the Habitat virtual environment, which allowed us to use simulation experiments to pre-train the model and then upload it to a real robot. Our architecture is adapted to work in a real-time environment and fully implements modern trends in this area.

show abstract

“…This paper presents a solution to these challenges by introducing VisionBlender, a synthetic dataset generator that operates with Blender. 1 Blender is increasingly becoming the premier solution for generating large training datasets (Veldhuizen 2018;Haim et al 2018;Villegas et al 2018;Newell and Densg 2020;Ranjan et al 2020;Alexopoulos et al 2020;Chen et al 2020aChen et al , 2020b. To the best of our knowledge, VisionBlender is the first user interface to facilitate the generation of endoscopic training datasets within Blender.…”

Section: Introductionmentioning

confidence: 99%

VisionBlender: a tool to efficiently generate computer vision datasets for robotic surgery

Cartucho

Tukra

Liu

et al. 2020

Computer Methods in Biomechanics and Biomedical Engineering: Im

View full text Add to dashboard Cite

BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation

Cited by 470 publications

References 10 publications

An Instance Segmentation-Based Method to Obtain the Leaf Age and Plant Centre of Weeds in Complex Field Environments

An Instance Segmentation-Based Method to Obtain the Leaf Age and Plant Centre of Weeds in Complex Field Environments

Real-Time Object Navigation With Deep Neural Networks and Hierarchical Reinforcement Learning

VisionBlender: a tool to efficiently generate computer vision datasets for robotic surgery

Contact Info

Product

Resources

About