Poly-YOLO: higher speed, more precise detection and instance segmentation for YOLOv3

Hurtík, Petr; Molek, Vojtěch; Hůla, Jan; Vajgl, Marek; Vlašánek, Pavel; Nejezchleba, Tomáš

doi:10.48550/arxiv.2005.13243

Cited by 8 publications

(11 citation statements)

References 31 publications

(52 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…3) Comparison with Single-stage object detection and Others: Poly YOLO [16] reported ∼ 22 fps on a 416×832 image with an AP score of 8.7. Other methods like Deep Watershed [48] and SGN [49] incur a huge computation complexity in their instance assignment techniques.…”

Section: State Of the Art Comparisonmentioning

confidence: 99%

See 1 more Smart Citation

Learning Panoptic Segmentation from Instance Contours

Chennupati¹,

Narayanan²,

Sistu³

et al. 2020

Preprint

View full text Add to dashboard Cite

Panoptic Segmentation aims to provide an understanding of background (stuff) and instances of objects (things) at a pixel level. It combines the separate tasks of semantic segmentation (pixel-level classification) and instance segmentation to build a single unified scene understanding task. Typically, panoptic segmentation is derived by combining semantic and instance segmentation tasks that are learned separately or jointly (multi-task networks). In general, instance segmentation networks are built by adding a foreground mask estimation layer on top of object detectors or using instance clustering methods that assign a pixel to an instance center. In this work, we present a fully convolution neural network that learns instance segmentation from semantic segmentation and instance contours (boundaries of things). Instance contours along with semantic segmentation yield a boundary-aware semantic segmentation of things. Connected component labeling on these results produces instance segmentation. We merge semantic and instance segmentation results to output panoptic segmentation. We evaluate our proposed method on the CityScapes dataset to demonstrate qualitative and quantitative performances along with several ablation studies.

show abstract

Section: State Of the Art Comparisonmentioning

confidence: 99%

“…Recently, single stage instance segmentation methods have been developed [15], [16]. These major approaches use fully convolution networks so that they can be trained in an endto-end fashion.…”

Section: Introductionmentioning

confidence: 99%

Learning Panoptic Segmentation from Instance Contours

Chennupati¹,

Narayanan²,

Sistu³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…A polygon is a generic representation for any arbitrary shape; however, it is more expensive to annotate than a bounding box. The object contour can be uniformly sampled in the range of 360°split into N equal polygon vertices, each represented by the radial distance r from the object's centroid as used in PolyYOLO [16]. We observe that uniform sampling cannot efficiently represent high curvature variations in the fisheye image object contours.…”

Section: B Generalized Object Detectionmentioning

confidence: 99%

OmniDet: Surround View Cameras Based Multi-Task Visual Perception Network for Autonomous Driving

Kumar

Yogamani

Rashed³

et al. 2021

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Surround View fisheye cameras are commonly deployed in automated driving for 360°near-field sensing around the vehicle. This work presents a multi-task visual perception network on unrectified fisheye images to enable the vehicle to sense its surrounding environment. It consists of six primary tasks necessary for an autonomous driving system: depth estimation, visual odometry, semantic segmentation, motion segmentation, object detection, and lens soiling detection. We demonstrate that the jointly trained model performs better than the respective single task versions. Our multi-task model has a shared encoder providing a significant computational advantage and has synergized decoders where tasks support each other. We propose a novel camera geometry based adaptation mechanism to encode the fisheye distortion model both at training and inference. This was crucial to enable training on the WoodScape dataset, comprised of data from different parts of the world collected by 12 different cameras mounted on three different cars with different intrinsics and viewpoints. Given that bounding boxes is not a good representation for distorted fisheye images, we also extend object detection to use a polygon with non-uniformly sampled vertices. We additionally evaluate our model on standard automotive datasets, namely KITTI and Cityscapes. We obtain the state-of-the-art results on KITTI for depth estimation and pose estimation tasks and competitive performance on the other tasks. We perform extensive ablation studies on various architecture choices and task weighting methodologies. A short video at https://youtu.be/xbSjZ5OfPes provides qualitative results.

show abstract

“…Subsequently, BshapeNet [26] designs extended frameworks by adding a bounding box mask branch providing additional information of object positions and coordinates to a Faster R-CNN to enhance the performance of instance segmentation. Recently, a few works propose some novel contourbased approaches for real-time instance segmentation [27], [28]. Deep Snake [27] uses the circular convolution for feature learning on the contour and proposes a two-stage pipeline including initial contour proposal and contour deformation for instance segmentation.…”

Section: B Instance Segmentationmentioning

confidence: 99%

“…Deep Snake [27] uses the circular convolution for feature learning on the contour and proposes a two-stage pipeline including initial contour proposal and contour deformation for instance segmentation. POLY-YOLO [28] increases the detection accuracy of YOLOv3 and realizes instance segmentation using tight polygon-based contour. However, most of these methods focus on appearance cues and can only segment objects that have been labeled as a specific category in a training set.…”

Section: B Instance Segmentationmentioning

confidence: 99%

U2-ONet: A Two-level Nested Octave U-structure with Multiscale Attention Mechanism for Moving Instances Segmentation

Wang¹,

Li²,

Luo

2020

Preprint

View full text Add to dashboard Cite

Most scenes in practical applications are dynamic scenes containing moving objects, so segmenting accurately moving objects is crucial for many computer vision applications. In order to efficiently segment out all moving objects in the scene, regardless of whether the object has a predefined semantic label, we propose a two-level nested Octave U-structure network with a multiscale attention mechanism called U 2 -ONet. Each stage of U 2 -ONet is filled with our newly designed Octave ReSidual U-block (ORSU) to enhance the ability to obtain more context information at different scales while reducing spatial redundancy of feature maps. In order to efficiently train our multi-scale deep network, we introduce a hierarchical training supervision strategy that calculates the loss at each level while adding a knowledge matching loss to keep the optimization consistency. Experimental results show that our method achieves state-of-theart performance in several general moving objects segmentation datasets.

show abstract

Poly-YOLO: higher speed, more precise detection and instance segmentation for YOLOv3

Cited by 8 publications

References 31 publications

Learning Panoptic Segmentation from Instance Contours

Learning Panoptic Segmentation from Instance Contours

OmniDet: Surround View Cameras Based Multi-Task Visual Perception Network for Autonomous Driving

U2-ONet: A Two-level Nested Octave U-structure with Multiscale Attention Mechanism for Moving Instances Segmentation

Contact Info

Product

Resources

About